overlay of flexible pavements: an ann approachethesis.nitrkl.ac.in/6190/1/e-32.pdf · overlay of...
TRANSCRIPT
Overlay of Flexible Pavements: An ANN Approach
A Report Submitted in Partial Fulfilment
of the Requirements for the degree of
Bachelor of Technology
in
Civil Engineering
by
Sonal Gumansingh Roll No.-110CE0069
Under Guidance of Prof. Mahabir Panda
Department of Civil Engineering
National Institute of Technology Rourkela
Rourkela-769008, Orissa, India
May, 2014
Overlay of Flexible Pavements: An ANN Approach
A Report Submitted in Partial Fulfilment
of the Requirements for the degree of
Bachelor of Technology
in
Civil Engineering
by
Sonal Gumansingh Roll No.-110CE0069
Under Guidance of Prof. Mahabir Panda
Department of Civil Engineering
National Institute of Technology Rourkela
Rourkela-769008, Orissa, India
May, 2014
CERTIFICATE
This is to certify that the thesis entitled, “Overlay of Flexible
Pavements: An ANN Approach” submitted by Miss. Sonal Gumansingh in
partial fulfilment of the requirement for the award of Bachelor of Technology
Degree in Civil Engineering at the National Institute of Technology, Rourkela
(Deemed University) is an authentic work carried out by her under my
supervision and guidance.
To the best of my knowledge, the matter embodied in the thesis has not
been submitted to any other University/ Institute for the award of any degree
or diploma.
Date: Prof. Mahabir Panda
Place: Rourkela Professor
Dept of Civil Engineering
National Institute of Technology
Rourkela-769008
NATIONAL INSTITUTE OF TECHNOLOGY ROURKELA-769008,ODISHA INDIA
Acknowledgments
I would like to express my gratitude to my guide, Prof. M. Panda, for his
encouragement, advice, mentoring and research support throughout my studies.
His technical and editorial advice was essential for the completion of this
dissertation. His ability to teach, depth of knowledge and ability to achieve
perfection will always be my inspiration.
My sincere thanks to Dr. S. K. Sarangi, Director and Prof. N. Roy, Head of
the Civil Engineering Department, National Institute of Technology Rourkela, for
his advice and providing necessary facility for my work.
I am very thankful to all the faculty members and staffs of civil engineering
department who assisted me in my research.
I would like to thank mtech. senior Aditya Kumar Das for his help during
final compilation of project report. I also thank all my batch mates, who have
directly or indirectly helped me in my project work and in the completion of this
report.
I am grateful to my parents Mr. Gadadhar Gumansingh and Mrs. Sarojshree
Gumansingh for their love, support and guidance. They have always been
supportive of my academic pursuit.
Sonal Gumansingh
Contents CERTIFICATE .................................................................................................................................................... 3
ABSTRACT ...................................................................................................................................................................... 7
List of Figures ................................................................................................................................................................. 8
List of Tables ................................................................................................................................................................ 10
ABBREVIATIONS ........................................................................................................................................................... 11
CHAPTER-1 ................................................................................................................................................................... 12
INTRODUCTION ........................................................................................................................................................... 12
1.1 GENERAL ............................................................................................................................................................ 12
1.2 FLEXIBLE PAVEMENTS ....................................................................................................................................... 13
1.3 CRITICAL RESPONSES FOR FLEXIBLE PAVEMENTS AND PREDICTION OF LIFE ......................................................... 13
1.4 SCOPE OF THE STUDY ........................................................................................................................................ 14
Chapter-2 ..................................................................................................................................................................... 15
LITERATURE REVIEW AND SCOPE OF THE STUDY ................................................................................................................... 15
2.1 LITERATURE REVIEW ......................................................................................................................................... 15
2.2 OBJECTIVE OF THE STUDY ................................................................................................................................ 16
Chapter 3 ..................................................................................................................................................................... 17
THEORY ........................................................................................................................................................................ 17
3.1 PAVEMENT STRUCTURE MODEL AND EVALUATION ......................................................................................... 17
3.2 OVERLAY DESIGN PROCEDURE....................................................................................................................................... 18
3.2.1 Modelling Pavement Response.................................................................................................................. 18
3.3 DIFFERENT PARAMETERS OF PAVEMENT DESIGN ...................................................................................................... 19
3.3.1 Effect of Temperature on Bituminous Layer Modulus ............................................................................... 19
3.3.2 Effect of Seasonal Variation ....................................................................................................................... 19
3.3.3 Effect of Poisson’s ratio ............................................................................................................................. 20
3. 4 DEVELOPED COMPUTER APPLICATIONS ...................................................................................................................... 21
3.5 ARTIFICIAL NEURAL NETWORK ...................................................................................................................................... 21
Chapter-4 ...................................................................................................................................................... 23
ANALYSIS METHODOLOGY ......................................................................................................................................................... 23
4.1 DESIGN-CASE DATABASE DEVELOPMENT ......................................................................................................... 23
4.2 TRAINING AND TESTING PROCESS FOR ACCURACY OF INPUT DATA ................................................................ 25
4.3 SENSITIVITY ANALYSIS ON TRAINED ANN .......................................................................................................... 26
4.3.1 Levenberg-Marquardt Backpropagation (LM) ........................................................................................... 26
4.3.2 Scaled Conjugate Gradient Backpropagation (SCG) ................................................................................... 27
4.3.3 Radial Basis Function (RBF) ........................................................................................................................ 27
4.4 TESTING PROCESS ON ACTUAL DATA OBTAINED FROM MORTH RESEARCH SCHEME R-81 .............................. 27
Chapter-5 ..................................................................................................................................................................... 29
ANALYSIS AND RESULTS ............................................................................................................................................... 29
5.1 DESIGNCASE DATABASE DEVELOPMENT........................................................................................................... 29
5.2 TRAINING AND TESTING PROCESS FOR ACCURACY OF INPUT DATA ........................................................... 29
5.3 SENSITIVITY ANALYSIS ON TRAINED ANN..................................................................................................................... 30
5.4 TESTING PROCESS ON ACTUAL DATA OBTAINED FROM MORTH RESEARCH SCHEME R-81 .......................... 35
5.4.1 THIN OLD PAVEMENTS ................................................................................................................. 37
5.4.2 THICK OLD PAVEMENTS ............................................................................................................... 42
5.4.3 EVALUATION OF COLD-MIX RECYCLED PAVEMENT SECTION ........................................................... 46
5.4.4 EFFECT OF TEMPERATURE ON PAVEMENT LAYER MODULI .............................................................. 48
SUMMARY ................................................................................................................................................................... 50
FUTURE SCOPE OF WORK ............................................................................................................................................ 50
REFERENCES ................................................................................................................................................................. 51
Appendix-A .................................................................................................................................................................. 53
A-1 EXAMPLE OF WINFLEX OUTPUT FILE ............................................................................................................................ 53
A-2 EXAMPLE OF TRAINED ANN DATASET ALONG WITH PERFORMANCE, TRAINING SET AND REGRESSION PLOT .................. 57
Appendix B ................................................................................................................................................................... 58
RADIAL BASIS FUNCTION DATA UNDER ANN FOR DIFFERENT DESIGNCASES ............................................................... 58
Appendix-C .................................................................................................................................................................. 61
FIELD EVALUATION OF PAVEMENT USING FWD .................................................................................................................... 61
C-1 THIN PAVEMENTS ............................................................................................................................................. 61
C-2 THICK PAVEMENTS ............................................................................................................................................ 66
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 7
ABSTRACT
Highway Pavements comprise of two types, flexible and rigid. In India Most of the roads
are constructed as flexible pavements. The main problem in flexible pavement is deterioration
due to traffic loading, material related factors and adverse climatic conditions. In order to avoid
and mitigate such difficulties, maintenance is to be done instead of reconstruction. Among all the
maintenance programs, the most common method adopted in India is to go for an asphalt overlay
on the old surface to increase the serviceability of the existing road. Hence, designing and
constructing the flexible overlay is very important regarding its performance. Designing an
overlay is challenging given restricted boundary conditions that must be observed and designed
for. Although, there is provided design code but some difficulties in solving process such as
accurate field data, error prone design curve reading, less accurate conversion formula for
temperature variation, time consuming calculations make it complex and dull to be used for
everyday purpose. In addition, collection and use the necessary data in the HMA overlay design
process needs spending a large amount of money and time and also the reliability and
comprehensive data. Unavailability of design software leads to manual calculation which is
prone to errors. Hence process should be implemented using a computer model to overcome
complexity, recurring tasks and time consuming method. An artificial neural network approach
can be used for the elimination of this drawback.
This study presents an attempt to apply artificial neural network to recommend asphalt
overlay thickness (HMA). Though noted common methods need time, reliable and some
essential data to be able to start designing process but artificial intelligence especially artificial
neural network is a method based on learning process which can find possible relation between
input and output sample data and is able to predict the output without any time with founded
relation quickly. Results of this study reveal that artificial neural network is appropriate for
implementation in predicting flexible overlay thickness.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 8
List of Figures Figure 1-1 Typical cross-section of a flexible pavement system (COURAGE, 1999) .................................................... 13
Figure 1-2 types of failure mechanisms in pavements (COURAGE,1999) ................................................................... 14
Figure 4-1 Multi layered Pavement Cross-section (Mostafa, 2009) ............................................................................ 23
Figure 4-2 Methodology Used (Mostafa, 2009) .......................................................................................................... 24
Figure 4-3 Data range of Input Values (Mostafa,2009) ............................................................................................... 25
Figure 5-1 WINFLEX Program (WINFLEX, 2006) ........................................................................................................... 28
Figure5-2 Graph of Percent Accuracy Vs No. of Hidden Nodes for 3layered pavement: Design-case 2 .................... 29
Figure 5-3 Graph of Percent Accuracy Vs No. of Hidden Nodes for 3layered Pavement: Design-case 3 .................... 29
Figure 5-4 Graph of Percent Accuracy Vs No. of Hidden Nodes for 3layered Pavement: Design-case 4 .................... 30
Figure 5-5 Graph of Percent Accuracy Vs No. of Hidden Nodes for 3layered Pavement: Design-case 5 .................... 30
Figure 5-6 Graph of Percent Accuracy Vs No. of Hidden Nodes for 3layered Pavement: Design-case 6 .................... 30
Figure 5-7 Graph of Percent Accuracy Vs No. of Hidden Nodes for 3layered Pavement: Design-case 7 .................... 30
Figure 5-8 Graph of Percent Accuracy Vs No. of Hidden Nodes for 4layered Pavement: Design-case 2 .................... 30
Figure 5-9 Graph of Percent Accuracy Vs No. of Hidden Nodes for 4layered Pavement: Design-case 3 .................... 30
Figure 5-10 Graph of Percent Accuracy Vs No. of Hidden Nodes for 4layered Pavement: Design-case 4 .................. 31
Figure 5-11 Graph of Percent Accuracy Vs No. of Hidden Nodes for 4layered Pavement: Design-case 5 .................. 31
Figure 5-12 Graph of Percent Accuracy Vs No. of Hidden Nodes for 4layered Pavement: Design-case 6 .................. 31
Figure 5-13 Graph of Percent Accuracy Vs No. of Hidden Nodes for 4layered Pavement: Design-case 7 .................. 31
Figure 5-14 Graph of Percent Accuracy Vs No. of Hidden Nodes for 5layered Pavement: Design-case 2 .................. 31
Figure 5-15 Graph of Percent Accuracy Vs No. of Hidden Nodes for 5layered Pavement: Design-case 3 .................. 31
Figure 5-16 Graph of Percent Accuracy Vs No. of Hidden Nodes for 5layered Pavement: Design-case 4 .................. 32
Figure 5-17 Graph of Percent Accuracy Vs No. of Hidden Nodes for 5layered Pavement: Design-case 5 .................. 32
Figure 5-18 Graph of Percent Accuracy Vs No. of Hidden Nodes for 5layered Pavement: Design-case 6 .................. 32
Figure 5-19 Graph of Percent Accuracy Vs No. of Hidden Nodes for 5layered Pavement: Design-case 7 .................. 32
A-1 Matlab screen shot for THICK OLD Km 123.795 to 124.000 of NH 6 .................................................................... 56
A-2 Performance Data ................................................................................................................................................. 56
A-3 Training State Plot ................................................................................................................................................. 56
A-4 Regression Plot ...................................................................................................................................................... 56
B-1 RBF data for 3layered pavement: Design-case 2 ................................................................................................... 57
B-2 RBF data for 3layered Pavement: Design-case 3 ................................................................................................... 57
B-3 RBF data for 3layered Pavement: Design-case 5 ................................................................................................... 57
B-4 RBF data for 3layered Pavement: Design-case 6 ................................................................................................... 58
B-5 RBF data for 3 layered Pavement: Design-case 7 .................................................................................................. 58
B-6 RBF data for 4layered pavement: Design-case 2 ................................................................................................... 58
B-7 RBF data for 4layered Pavement: Design-case 3 ................................................................................................... 59
B-8 RBF data for 4layered Pavement: Design-case 5 ................................................................................................... 59
B-9 RBF data for 4 layered Pavement: Design-case 7 .................................................................................................. 59
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 9
List of Tables
Table 5-1 Transfer Function for 3layered pavement Design...................................………………………..................………33
Table 5-2 Transfer function for 4 layered pavement Design………………..………………………………….……..............………..34
Table 5-3 Transfer function for 5 layered pavement Design…………………………………………..……….…..............….……….34
Table 5-4 Details of selected test sections (MORTH R-81)..........................................................................................35
Table 5-5 Modulus Range (MPa) for types of surfacing (MORTH R-81)......................................................................36
Table 5-6 CASE Thin_1: ANN data for Km 1.865 to 2.000 of SH * (Salua Road) for the Deflection Data Collected during the Year 2001-02……………………………………………………………………………………………………..................................….37
Table 5-7 CASE Thin_2: ANN data for Km 2.850 to 3.000 of SH* (Salua Road) for the Deflection Data Collected during the Year 2001-02………………………………………………………………………………………………................................……...…37
Table 5-8 Case Thin_3:ANN data for Km 4.625 to 5.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-02…………………………………………..……………………………………………………….............................................….38
Table 5-9 CASE Thin_4: ANN data for Km 3.370 to 4.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-02…………………………………………….………………………………………………………...........................................…39
Table 5-10 CASE Thin_5: ANN data for Km 15.000 to 15.270 of NH-60 for the Deflection Data Collected during the Year 2001-02………………………………………………..………………………………………………………………….................................….40
Table 5-11 CASE Thick_1: ANN data for Km 123.795 to 124.000 of NH-6 for the Deflection Data collected during the Year 2000-01 (Kumar, 2001)…………………………………………………………………………….................................………………….41
Table 5-12 CASE Thick_2: ANN data for Km 125.000 to 125.270 of NH-6 for the Deflection Data collected during Summer Season of the Year 2001-02………………………………………….…………………………...........................………………….42
Table 5-13 CASE Thick_3: ANN Data for Km 134.000 to 134.270 of NH-6 for the Deflection Data collected during the Year 2001-2002……………………………………………….………………………………………………………..................................………….43
Table 5-14 CASE Thick_4: ANN data for Km 150.000 to 150.245 of NH-6 for the Deflection Data collected during the Year 2001-02……………………………………………………………………………………………………………………..................................….44
Table 5-15 CASE Thick_5: ANN data for Km 151.000 to 151.245 of NH-6 for the Deflection Data collected during the Year 2001-02…………………………………………………………………………………………………………………...................................…....44
Table 5-16 CASE Thick_7: ANN data for Km 152.000 to 152.245 of NH-6 for the Deflection Data collected during the Year 2001-02…………………………………………………………………………………………………………………...................................…....45
Table 5-17 CASE Thick_8: ANN data for Km 153.000 to 153.245 on NH-6 for the Deflection Data collected during the Year 2001-02………………………………………………………………………………………………………………..........................……….46
Table 5-18 back calculated layer Moduli Range for Recycled Pavement Stretch(MORTH R-81)................................46
Table 5-19 ANN data for Cold-mix Recycled pavement Stretch………………………………………………….………….................46 Table 5-20 Layer Moduli for the Deflection Data collected on KM 112. 000 to 112.540 of NH-6 at Different Pavement Temperatures (MORTH R-81)....................................................................................................................47 Table 5-21 ANN data for effect of temperature on pavement section…………………………………….................……………50 Table C-1 CASE Thin_1: Km 1.865 to 2.000 of SH * (Salua Road) for the Deflection Data Collected during the Year
2001-02. (MORTH Research Scheme R-81) ................................................................................................................. 60
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 10
Table C-2 Thickness Details for the Stretch from Km 1.820 to 2.000 of SH*(SALUA Road) (MORTH Research Scheme
R-81) ............................................................................................................................................................................ 60
Table C-3 CASE Thin_2: Km 2.850 to 3.000 of SH* (Salua Road) for the Deflection Data Collected during the Year
2001-02. (MORTH Research Scheme R-81) ................................................................................................................. 61
Table C-4 Thickness Details for the Stretch from Km 2.850 to 3.000 of SH (SALUA Road) (MORTH Research Scheme
R-81) ............................................................................................................................................................................ 61
Table C-5 Case Thin_3: Km 4.625 to 5.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-
02. (MORTH Research Scheme R-81)........................................................................................................................... 62
Table C-6Thickness Details for the Stretch from Km 4.625 to 5.000 of SH (IIT Bypass) (MORTH Research Scheme R-
81) ................................................................................................................................................................................ 62
Table C-7 CASE Thin_4: Km 3.370 to 4.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-
02 (MORTH Research Scheme R-81) ............................................................................................................................ 63
Table C-8 Thickness Details for the Stretch from Km 3.370 to 4.000 of SH (IIT Bypass) (MORTH Research Scheme R-
81) ................................................................................................................................................................................ 63
Table C-9 CASE 5: Km 15.000 to 15.270 of NH-60 for the Deflection Data Collected during the Year 2001-02
(MORTH Research Scheme R-81) ................................................................................................................................ 64
Table C-10 Thickness Details for the Stretch from Km 15.000 to 15.270 of NH-60 (MORTH Research Scheme R-81)64
Table C-11 CASE Thick_1: Km 123.795 to 124.000 of NH-6 for the Deflection Data collected during the Year 2000-01
(Kumar, 2001) (MORTH Research Scheme R-81) ......................................................................................................... 65
Table C-12 Thickness Details for the Stretch from Km 123.795 to 124.000 of NH-6 (MORTH Research Scheme R-81)
..................................................................................................................................................................................... 65
Table C-13 CASE Thick_2: Km 125.000 to 125.270 of NH-6 for the Deflection Data collected during Summer Season
of the Year 2001-02. (MORTH Research Scheme R-81) ............................................................................................... 66
Table C-14 Thickness Details for the Stretch from Km 125.000 to 125.270 of NH-6 (MORTH Research Scheme R-81)
..................................................................................................................................................................................... 66
Table C-15 CASE Thick_3: Km 134.000 to 134.270 of NH-6 for the Deflection Data collected during the Year 2001-
2002. (MORTH Research Scheme R-81)....................................................................................................................... 67
Table C-16 Thickness Details for the Stretch from Km 134.000 to 134.270 of NH-6 (MORTH Research Scheme R-81)
..................................................................................................................................................................................... 67
Table C-17 CASE Thick_4: Layer Moduli for Km 150.000 to 150.245 of NH-6 for the Deflection Data collected during
the Year 2001-02 (MORTH Research Scheme R-81) .................................................................................................... 68
Table C-18 Thickness Details for the Stretch from Km 150.000 to 150.245 of NH-6 (MORTH Research Scheme R-81)
..................................................................................................................................................................................... 68
Table C-19 CASE Thick_5: Km 151.000 to 151.245 of NH-6 for the Deflection Data collected during the Year 2001-02
(MORTH Research Scheme R-81) ................................................................................................................................ 69
Table C-20 Thickness Details for the Stretch from Km 151.000 to 151.245 of NH-6 (MORTH Research Scheme R-81)
..................................................................................................................................................................................... 69
Table C-21 CASE Thick_7: Km 152.000 to 152.245 of NH-6 for the Deflection Data collected during the Year 2001-02
(MORTH Research Scheme R-81) ................................................................................................................................ 70
Table C-22 Thickness Details for the Stretch from Km 152.000 to 152.245 of NH-6 (MORTH Research Scheme R-81)
..................................................................................................................................................................................... 70
Table C-23 CASE Thick_8: Km 153.000 to 153.245 on NH-6 for the Deflection Data collected during the Year 2001-
02 (MORTH Research Scheme R-81) ............................................................................................................................ 71
Table C-24 Thickness Details for the Stretch from Km 153.000 to 153.245 of NH-6 (MORTH Research Scheme R-81)
..................................................................................................................................................................................... 71
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 11
ABBREVIATIONS SYMBOL
E1 : Elastic modulus of A.C. layer.
E2 : Elastic modulus of granular base layer.
E3 : Elastic modulus of subgrade.
f1, f2, f3 : Coefficients of fatigue criterion.
f4, f5 : Coefficients of permanent deformation criterion.
Nƒ : allowable number of load repetitions to prevent fatigue cracking.
Nd : maximum allowable number of load repetitions to limit permanent deformation,
Mr : Resilient modulus.
p : Mean normal stress (σ1 + 2 σ3) / 3.
P : Concentrated load.
εc : Vertical compressive strain on the surface of subgrade.
εr : Radial strain or recoverable strain.
εt : Tensile strain at bottom of asphalt layer.
εz : Vertical strain.
σ1,2,3 : Three principal stresses.
σz : Vertical stress.
τ : Shear stress.
μ : Poisson's ratio.
ANN: Artificial Neural Network
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 12
CHAPTER-1
INTRODUCTION
1.1 GENERAL
Pavements are of two types, flexible and rigid. In flexible layers, the common ones are
Asphaltic wearing course above base and/or sub-base which rests on subgrade layer. Generally,
failure of pavement occur due to factors affecting serviceability such as traffic loading and
weather conditions. For preservation, Maintenance Rehabilitation (MR) activities should be done
in well-timed manner. The most popular M and R activities is use of Hot Mix Asphalt (HMA)
overlays. This improves the roadway by utilizing the existing layers of asphalt concrete and
aggregate base as support for a new wearing surface. Adding a defined thickness of asphalt
concrete provides structural reinforcement to the roadways, permits the designer to allow change
in traffic use, addresses increasing traffic volume and weights, corrects riding qualities, and
efficiently extends structural performance of the roadway for ten or even twenty years while
maintaining factors like the structural capacity of each layer, and the variation of material
properties due to seasonal and environmental changes, especially the temperature and moisture
variations. A significant consideration during design and ultimately construction of overlay is the
potential for existing pavement cracking of different severity to reflect through new wearing
surface, with obvious aesthetic impacts. HMA overlay provides a reasonably fast, lucrative
means of restoring satisfaction of user, rectifying deficiencies of existing surface, and promoting
structural load-carrying capacity reckoning on the designed thickness.
Indian Road Congress (IRC) IRC : 81 – 1997 uses design curves relating characteristic
pavement deflection to cumulative standard axle loads. Thus thickness of Bituminous Macadam
has to be manually calculated which is time consuming and prone to error by user. Hence, it
should be carried out in a computer environment to overcome complex and time consuming
procedure. This study incorporates the practicability of application of ANN approach for
recommendation of proper pavement overlay thickness based on knowledge from overlay design
cases. For developing countries like India where dedicated software for the purpose are
unavailable, this approach is productive in every manner.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 13
1.2 FLEXIBLE PAVEMENTS
The main structural function of a pavement is to support the induced load by traffic and
to distribute these loads safely to the foundation. Figure 1.1 shows the typical cross-section of a
flexible pavement system. This pavement comprises a number of bituminous layers placed over
the road base (unbound or bound material) over a similar unbound sub-base material placed on
the natural subgrade. This pavement is referred to as "flexible" because the bituminous materials
are capable of flexing slightly under traffic loading. For thinly surfaced pavements, the road base
is often unbound granular material. The base course immediately beneath the surface course can
be composed of crushed stone, crushed slag or other untreated or stabilized materials. The sub-
base course is the layer beneath the base course. The reason that two different granular materials
are used is for economy. Instead of using the more expensive base course material for the entire
layer, local and cheaper materials can be used as a sub-base course on top of subgrade. (Huang,
2004), (Witczack and Yoder, 1975).
Figure 1-1 Typical cross-section of a flexible pavement system (COURAGE, 1999)
1.3 CRITICAL RESPONSES FOR FLEXIBLE PAVEMENTS AND PREDICTION OF LIFE
Kerkhoven and Dormon (1953) first suggested the use of vertical compressive strain on
the surface of the subgrade as a failure criterion to reduce the permanent deformation of
subgrade which is the main reason of subgrade rutting. After that Saal and Pell (1960)
recommended the use of horizontal tensile strain at the bottom of the asphaltic concrete to
minimize fatigue cracking which limits the fatigue life of pavements(Huang, 2004). Two main
structural failure mechanisms for flexible pavements which are the fatigue cracking of the
asphaltic concrete layer and permanent deformation of the subgrade (rutting) are also shown in
figure 1-2.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 14
Figure 1-2 types of failure mechanisms in pavements (COURAGE,1999)
1.4 SCOPE OF THE STUDY
During the last three decades the approach to design of flexible pavements has begun to
undergo transformation from empirical method to mechanistic method because of the improved
understanding of the behaviour of materials and the availability of analytical tools for the
analysis of pavements. In India also, a massive pavement performance study was undertaken
during 1983 to 1993 [Research Schemes R-6, 1995; R-19 and R-56, 1999] as a result of which a
new standard for design of flexible pavements was published by Indian Roads Congress [IRC:
37, 2001]. The new design method uses a mechanistic approach for the determination of design
pavement thicknesses.
The main concern among the researchers in India in using the empirical relationships
recommended by IRC is that there has not been any validation of the relationships for the
specifications and construction practices adopted in India. Thus, it is essential to have adequate
data for selection of realistic layer moduli appropriate for the conditions prevailing in India.
Hence, a rational approach for prediction is desirable for the analytical design of pavements and
overlays. Also Extensive calculation and time consuming design curve readings prone to reading
error by user, makes road for alternative method which is less time consuming and efficient.
Hence Artificial Neural Network finds its use in predicting overlay thickness from the learned
relationship through repeated presentation of data and is capable of generalizing to new and
previously unseen data. In using ANN though initially needs large amount of input data for
learning database; it is quick at finding output for new data.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 15
Chapter-2
LITERATURE REVIEW AND SCOPE OF THE STUDY
2.1 LITERATURE REVIEW
ANNs are helpful computation tools increasingly used for solving complex, resource-
concentrated problems as an alternative to using traditional techniques. Thube et al. (2006) used
ANNs for forecasting cracking, ravelling, rutting and roughness for Low Volume Roads (LVR)
in India. This pavement condition prediction methodology using ANN used Road inventory
data, as well as 6cycles of pavement performance data from in-service LVR pavement sections
over a 3 year period in India. Ceylan et al. (2004) used ANNs as pavement structural analysis
tools to be compared with calibrated HDM-4 models developed by World Bank. Meier et al.
(1997) trained ANNs as surrogate to ELP analysis for back calculating pavement layer moduli
and found a 42 times increase in processing speed for prediction of critical responses and
deflection profiles of flexible pavements. Gucunski and Krstic (1996) and Khazanovich and
Roesler (1997) reported similar ANN applications.
The research project team working on the development of the new, mechanistic based
AASHTO Pavement Design (NCHRP 1-37A) has recognized ANN as non-traditional yet
possessing very powerful computing techniques, taking advantage of ANN models in preparing
the 2002 Design Guide concrete pavement analysis package. Gholamali et al.(2013) proposed
an AASHTO revised method with the goodness of ANNs to reduce the time and the calculation
procedure for selection of the asphalt overlay thickness (HMA) by back propagation network
with 1-13-3 combination as the optimum network to estimate the overlay thickness.
Retherford(2012) developed a systematic uncertainty management surrogate model with
objective of addressing model uncertainty, reducing computational expense, incorporating
uncertainty in pavement design for risk-based mechanistic-empirical (M-E) pavement design.
Mahoney et al.(1989) integrated non-destructive testing (NDT) for estimating layer
resilient moduli, seasonal moduli adjustment, and failure criteria of asphalt concrete using
EVERPAVE for Washington state flexible pavements. Terzi(2005) modelled ANN model for
serviceability index of flexible pavement with higher regression value than AASHTO.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 16
Gopalkrishnan(2010) discoursed the effect of training algorithm on neural networks when it is
aiding pavement design.
Abdallah et al, (2000) developed ANN models for the determination of critical strains at
layer interfaces for 3 and 4 layered flexible pavement systems. It could predict remaining life and
required overlay thickness for an existing pavement using fatigue and rutting criteria. Ceylan et
al.(2007), Gopalakrishnan and Thompson(2004) worked on FE solutions database to be used for
predicting critical pavement response and layer moduli when incorporated with ANNs.
In addition, artificial neural networks (Attoh-Okine, 1994, 1999, 2001, 2002 ; Choi,
Adams, and Bahia, 2004; Sundin and Braban-Ledoux, 2001; Roberts and Attoh-Okine, 1999;
Alsugair and Al-Qudrah, 1998; Huang and Moore, 1997; Fwa and Chan, 1993) have recently
been used in flexible pavement cracking prediction, condition ratings of jointed concrete
pavements, pavement-performance prediction and simulating pavement deterioration. Several
neural network studies, as explained above, have been conducted to assist engineers in selecting
optimal maintenance and rehabilitation activities.
2.2 OBJECTIVE OF THE STUDY
The main objectives of the study are
a. To apply artificial neural network for recommending pavement overlay thickness based
on learning from overlay design cases avoiding complicated and time consuming input
file preparation, complexity procedures and repeated tasks under IRC Pavement Design
Guide.
b. To bring forward advanced intelligent learning techniques for pavement designs into
practical use.
c. To compare and find accuracy rate of predicted overlay thickness from ANN with
calculated ones determined by the MORTH research scheme-81 attempting to use proper
transfer function for better accuracy in definite time.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 17
Chapter 3
THEORY
3.1 PAVEMENT STRUCTURE MODEL AND EVALUATION
The pavement is regarded as a multi-layered elastic system. The application of the
multilayer elastic theory in pavement analysis involves several assumptions: (Burmister, 1943)
1. Material in the pavement layers is assumed to be elastic, homogenous and isotropic.
2. All layers, except the bottom one, are finite in depth and the bottom layer (subgrade) is
assumed to be infinite depth.
3. All layers are assumed to be infinite in extent in the lateral direction.
4. The applied loads are static and load imprints are assumed to be circular.
5. Pavement materials are characterized by a modulus of elasticity and Poisson’s ratio.
6. Full friction is assumed to have developed between layers at each interface.
These assumptions may be found to be unrealistic for actual pavement conditions. However, the
existing mechanistic procedures showed that applying the multi-layered elastic theory gave
acceptable results. (Smith, 1986)
Structural evaluation of pavements commonly involves applying a standard load to the
pavement and measuring its response which may be stress, strain or deflection. The most
commonly measured response is deflection. Since its development in 1953, the Benkelman Beam
has become a standard tool used by several agencies for non-destructive testing of pavements.
Benkelman Beam is a 3.66 m long, portable instrument used to measure surface deflection of the
pavement loaded by the rear axle of a standard truck. The main disadvantage with this equipment
is that the support legs of the beam often lie within the deflection basin, which affects the
measured deflections. Also, a single deflection does not give adequate information about the
condition of various layers of the pavement. Hence Research scheme MORTH R-81 used
dynamic loading type Falling Weight Deflectometer (FWD) for the purpose of calculation of
deflection.
The basic working principle of the impulse loading equipment is to drop a mass on the
pavement to produce an impulse load and measure the surface deflections. The mass is dropped
on a spring system, which in turn transmits the load to the pavement through a loading plate. The
resulting deflection bowl characteristics are observed and used in the back-calculation of
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 18
pavement material properties. The two models developed at IIT Kharagpur in India [Kumar et al,
2001; Reddy et al, 2002a] can be listed in this category.
3.2 OVERLAY DESIGN PROCEDURE
In overlay design procedure, thicknesses are so calculated to put damages within the
allowable limits in the existing pavement and the new overlay. Parameters affecting overlay
thickness have been discussed in section 3.3. Diagrammatically it can be shown as follows:
3.2.1Modelling Pavement Response
Modelling pavement response is normally used for estimation of two most common
distress types found in AC pavements: fatigue cracking and rutting distress. Fatigue cracking is
caused by horizontal tensile strain at the bottom of asphalt layers, while rutting distress origin
from vertical compressive strain on the top of the subgrade layer.
The failure criterion for fatigue cracking is expressed as generally;
Nf = f1 (Єt)-f2
(E1)-f3
[Huang, 2004]
Research Scheme MORTH R-81 uses 2.21E-4 , 3.89 , 0.854 for f1,f2,f3 respectively.
Temperature: Asphalt Bound
Moisture: unbound Material
Rainfall
Environmental
Factor
Axle Loading
Traffic and
Loading
Condition
Analysis
Effect of Seasonal
Variation (Soil type,
Plasticity Index, Field
Moisture Content)
Deflection
Prediction
Material Properties of each layer
(Modulus of elasticity, Poisson’s
Ratio, Layer Thickness)
Back-calculation of
Deflection Data
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 19
The failure criterion for permanent deformation is expressed as;
Nd= f4 (Єc)-f5
[Huang, 2004]
As per Research Scheme MORTH R-81 values of ƒ4 and ƒ5 suggested as 4.1656E-8 and 4.4337.
3.3 DIFFERENT PARAMETERS OF PAVEMENT DESIGN
Parameters affecting pavement design also affects the selection of database for input preparation.
For effective input parameters which will give better accuracy it is needed to consider below
factors for its effect on moduli values and others.
3.3.1 Effect of Temperature on Bituminous Layer Modulus
Properties of bituminous mixes like stiffness get affected by change in temperature which
affects deflection. Modulus values determined at different temperatures are normally adjusted to
correspond to a standard temperature (35◦C) for design of pavements and overlays. Different
temperature adjustment factors and equations were given by various researchers for adjusting the
modulus and/or deflections for temperature. Code says, correction for deflection values
measured at pavement temperature other than 35◦C should be 0.01mm for each degree change
from standard temperature. WINFLEX uses the following relationship
En=Intercept-(Slope X T) (WINFLEX, 2006)
where, exponent n=0.35 and slope=0.12692
Examination of the deflections measured at different pavement temperatures indicates
that the variation of deflections with temperature is maximum in the case of deflections close to
the centre of the loading plate in NDT test. There is not much effect of temperature on the
deflections measured away from the load. It is widely recognized that deflections at locations
closer to the load are mostly dependent on the moduli of all the layers whereas defections, at
large distances depend on modulus of subgrade only. (Research Scheme MORTH R-81)
3.3.2 Effect of Seasonal Variation
Deflection depends upon change in climate. Worst climate (after Monsoon) affects the
pavement badly. It depends on subgrade soil and moisture content. Correction for seasonal
variation in code depends on types of soil subgrade (sandy/gravelly or Clayey with PI<15or
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 20
Clayey with PI>15), field moisture content, average annual rainfall (<1300mm or>1300mm). In
WINFLEX, climate zone 3 was considered which was nearer to hot climate zone.
Figure 3-1 IRC Seasonal correction factor (clayey subgrade, PI>15,rainfall<1300) (IRC: 81-1997)
The summer deflections were lower whereas became larger soon after the monsoon. Even
though the modulus of the bituminous bound layer is expected to be lower in summer when
temperatures are high, the subgrade and granular layers have higher elastic modulus in summer
resulting in lower overall deflections in case of thin bituminous surfacing layers. The behaviour
of cracked bituminous layers is close to that of granular layers and has high elastic modulus in
summer than the modulus obtained during the monsoon (MORTH R-81). Below is the graphical
representation of subgrade modulus variation throughout year as per WINFLEX, 2006. This is
closely similar to the one explained under MORTH R-81.
Figure 3-2 Graphical Representation of Subgrade Modulus Variations Throughout the year (WINFLEX,2006)
3.3.3 Effect of Poisson’s ratio
For most of the pavement materials, the influence of Poisson’s ratio (μ) on various
critical parameters is normally small. This allows the use of typical values in the analysis rather
than resorting to complex laboratory testing (Mitchell and Monismith, 1977).
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 21
According to IRC: 37-2001 Asphalt layer has Poisson’s ratio in the range of 0.35-0.5 while 0.5 is
for temperature range of 35◦C to 40
◦C and 0.35 for the range of 20
◦C to 30
◦C. The value off
Poisson’s ratio for granular layer and subgrade layer is 0.4.
3. 4 DEVELOPED COMPUTER APPLICATIONS
Mechanistic-based pavement design methods are being developed and used in different
countries such as Europe and North America. For example in the USA, the Washington State
Department of Transportation (WSDOT) uses ME design system in the computer program
EVERPAVE 5.0, March 1999 which was developed at the University of Washington. Bayomy et
al. (1996) developed WINFLEX under Idaho Department of Transportation (ITD) which also
uses the same design system for flexible pavements and implemented in the computer program
WINFLEX, May 2001. This program is available in internet and the link of which is provided in
the reference. Here WINFLEX calculates overlay thickness for input design-cases which were to
be checked with predicted ones for accuracy.
3.5 ARTIFICIAL NEURAL NETWORK
Neural networks are valuable intelligent tools which have been used significantly in
engineering applications when it is difficult to prosecute conventional methods showing inferior
performance (Fwa and Chan, 1993). Also Artificial Neural Networks (ANNs) have proved to
outperform traditional modelling counterparts in solving various complex engineering problems.
ANNs are intelligent systems based on uncomplicated computing models of biological structure
of the human brain (Venayagamoorthy et al, 2002). The highly connected, distributed nature of it
imparts high degree of fault tolerance, noise immunity, and generalization capability. It consists
of parallelly operating simple elements, inspired by biological nervous systems. The network
function of ANN is determined by the connections between elements which can be trained by
adjusting the values of the connections (weights) between the elements. Batch training of a
network proceeds by making weight and bias changes based on an entire set (batch) of input
vectors. It may also be referred to as ‘‘on line’’ or ‘‘adaptive’’ training.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 22
Figure 3-3 Multilayer feed-forward Neural network (Venayagamoorthy,2004)
Typically, in a Multilayer Feed-forward Neural Network (MLFNN), there is an input
layer, one or more intermediate layers called the hidden layers and an output layer, which
outputs the network’s response to its inputs. The activity of the input units represents the raw
information that is fed into the network, while that of each hidden unit is determined by the
activities of the input units and the weights on the connections between the input and the hidden
units. Behaviour of the output units depends on both. The different layers structure allows
flexibility, more information capturing and relationship identification between variables.
(Venayagamoorthy, 2004)
Figure 3-4 Basic Principle of Neural Network (Terzi, 2007)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 23
Chapter-4
ANALYSIS METHODOLOGY
For developing an ANN-based overlay design thickness procedure, required is database
of design cases. Hence broad analysis was done on suggested three typical flexible pavement
cross sections: 3-layer, 4-layer, and 5-layer, considering new overlay and the subgrade. For each
cross section, 7 failure cases to be controlled by the design have been selected creating 21 design
cases. A range of input data has to be suggested for each design case. WINFLEX computer
program, was used for calculating the overlay thickness. Consequently, design cases database
can be created. Training data sets were then being selected to be used in training process for
ANN approach. Matlab neural network toolbox was used in designing and training the neural
network. Sensitivity analysis on the trained ANN has been conducted. Finally, testing or
validating process was performed on the trained ANN. Figure 4-2 presents in details the analysis
methodology adopted in this study (Mostafa, 2009), where it is divided into the following four
main steps:
1. Design-case Database Development
2. Training and Testing process for accuracy of INPUT data
3. Sensitivity analysis on trained ANN
4. Testing Process on actual data obtained from MORTH R-81
4.1 DESIGN-CASE DATABASE DEVELOPMENT
For building design cases database, three different pavement cross sections have been
suggested: 3-layer, 4-layer, and 5-layer as shown in Figure. The information to be taken into
account are the material data inputs for each layer: elastic modulus (E), layer thickness (H), and
Poisson ratios (ν) as shown. The assumption for elastic layer system is previously discussed in
3.1.
Eov, Hov, νov New Overlay
Eov, Hov, νov New Overlay Eold, Hold, νold Old Asphalt
Eold, Hold, νold Old Asphalt Eb, Hb, νb Base
/////////// //////// //////// ////////// /////// ////////
Esg, νsg Subgrade Esg, νsg Subgrade
3 Layer 4 Layer
Eov, Hov, νov New Overlay
Eold, Hold, νold Old Asphalt
Eb, Hb, νb Base
Esb, Hsb, νsb Subbase
//////// //////// ///////////
Esg, νsg Subgrade
5 Layer
Figure 4-1 Multi layered Pavement Cross-section (Mostafa, 2009)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 24
START
Design Cases
3 Layer
4 Layer
5 Layer
Data Range
Cross
sections
7 controlled
Failure cases
IRC overlay design
Design Cases Database
R-81 Actual dataset
from backcalculation Training Sets
Other Training Sets Training Process
MATLAB neural
network toolbox
Sensitivity Analysis
Type of Transfer function
No. of Hidden Nodes
No. of Hidden Layers
Trained ANN
Testing Process with actual
R-81 data using specific
Transfer function found
using Sensitivity analysis
Accuracy
Rate
Low
Implement
High
It is noteworthy that the overlay thickness, Hov, is the desired output to be calculated.
Poisson ratios, performance prediction model for fatigue and rutting failure, and modulus-
temperature adjustment for AC layers have been assumed to be constant for every design-case.
Their values are discussed in theory section under 3.3.
For each run, the overlay design software, WINFLEX, determines overlay thickness for the 3
cross sections based on controlling of both fatigue failure at the bottom of the bituminous layers
and/or Rutting failure on the top of the subgrade layer. The fatigue failure mode is selected
according to three conditions as follow:
The first condition of fatigue failure is selected so that the design is controlled by
considering fatigue failure in the old Asphalt and the new overlay layer.
The second fatigue failure mode is selected so that only fatigue failure in the new
overlay layer is considered.
Figure 4-2 Methodology Used (Mostafa, 2009)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 25
The last condition is chosen with the purpose of fatigue failure only in the old AC layer.
As a result and account for rutting failure with/without, there are seven failure modes for each
cross section, i.e., total of 21 design cases, as follow:
Using the data range, design cases database can be created through more than 600 runs for
each databases. Range of data can be seen from the given picture (Mostafa, 2009)
Figure 4-3 Data range of Input Values (Mostafa,2009)
It is noteworthy that pavement is considered to be failed when the total damage reaches
100%, whether it is due to fatigue or rutting. If the calculated overlay thickness is equal to 2.54
mm, which was specified as increment value in WINFLEX, this means that there is no need for
overlay.
4.2 TRAINING AND TESTING PROCESS FOR ACCURACY OF INPUT DATA
In order to develop ANN-based overlay design approach, it must be well trained using
training sets extracted from the developed database. In this analysis, Input datas are
“crosssection”, “Failure Mode”, “Eov”, “Eold”, “Eb”, “Esb”, “Esg”, “Hold”, “Hb”, “Hsb”, ESAL
whereas the output is Hov. The number of hidden layers and transfer function should be specified.
Failure in the old AC layer and the new overlay layer
Considering fatigue failure in new overlay and old pavement.
Considering both rutting and fatigue failure in new overlay and
old pavement.
Failure in the new overlay layer only
Considering fatigue failure in new overlay.
Considering both rutting and fatigue failure in new overlay.
Failure in the old AC layer
Considering fatigue failure in old pavement.
Considering both rutting and fatigue failure in old pavement.
Rutting on the sub grade layer
Considering rutting on the sub grade layer
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 26
In this study, 2 hidden layers were used. Here requirement is to get best mapping for the input to
the desired output (Hov). Each input set produces output under ANN with varied random set of
initial weigts. By training the network, the weights of the system continually adjust to
incrementally reduce the difference between the output and the desired response. This difference
is cited as the error and can be measured in different ways. The most common measurement is
the Mean Squared Error (MSE). The MSE is the average of the squares of the difference between
each output and the desired output.
4.3 SENSITIVITY ANALYSIS ON TRAINED ANN The objective of the sensitivity analysis is to express the fitness of neural networks as an
effective way in calculating overlay thickness with the most achievable accuracy that can be
obtained for the most economical cost. The neural networks were influenced to several
parameters that guarantee the greatest achievable accuracy such as type of transfer function,
number of hidden nodes, and hidden layers. Gopalkrishnan (2010) tried to use different training
algorithm of which transfer functions that were considered for each design cases were
• TRAINLM-Lavenberg-Marquardt
• TRAINSCG-Scaled Conjugate Gradient
• Radial Basis Function
4.3.1 Levenberg-Marquardt Backpropagation (LM)
The LM second-order numerical optimization technique combines the advantages of
Gauss–Newton and steepest descent algorithms. While this method has better convergence
properties than the conventional back-propagation method, it requires O(N2) storage and
calculations of order O(N2) where N is the total number of weights in an MLP back-propagation.
This algorithm is considered to be very efficient when training networks which have up to a few
hundred weights. Although the computational requirements are much higher for each iteration of
the LM training algorithm, this is more than made up for by the increased efficiency. This is
especially true when high precision is required (MATLAB Toolbox, User’s Guide, 2010).
4.3.2 Scaled Conjugate Gradient Backpropagation (SCG)
The SCG training algorithm was developed to avoid this time-consuming line search,
thus significantly reducing the number of computations performed in each iteration, although it
may require more iteration to converge than the other conjugate gradient algorithms. The storage
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 27
requirements for the SCG algorithm are lesser than that of LM and it finds its use when no of
input data is very high. (MATLAB Toolbox, User’s Guide, 2010)
4.3.3 Radial Basis Function (RBF)
In MLP network weighing process of Input Variable is used while RBF gives equal
importance to all Input parameters. This makes it less important at calculation of overlay design
of flexible pavement as weighing process of different variables is important.
4.4 TESTING PROCESS ON ACTUAL DATA OBTAINED FROM MORTH RESEARCH SCHEME R-81
Once the network is trained, testing process should be started. The trained network
should be exposed to the data sets obtained from ministry of Road Transport and Highways
(Research scheme R-81) which gives falling weight deflectometer test result of pavements tested
by IIT kharagpur. Different design cases were selected to represent training data sets which were
taken as INPUT. The predicted overlay thickness using the trained network of ANN software
should be compared with the actual ones from MORTH R-81 to come up with the accuracy rate
or reliability. The transfer function to be used should be checked with sensitivity analysis result
for optimal function for given database. If the accuracy rate is low, then the network is not
properly trained and other training sets should be generated to retrain the network, otherwise, the
network is considered to be reliable and ready for implementation. (Mostafa, 2009)
The implementation is solely dependent on accuracy of data. The accuracy and time
taken to reach required accuracy are important in the sense of implementation in program for
everyday use. The accuracy is highly subjected to input data which though randomly taken yet
are in resonance with parameters affecting overlay thickness. If a set of input data doesn’t give
required accuracy in optimal number of runs, effort should be taken to change the dataset to keep
it in resonance with parameters affecting overlay design. Effect of parameters is discussed in
section 3.3. After getting accuracy in optimal number of runs a set of input can be universally
used for the same design-case purpose at different instances i.e. it can be available for use in
ANN based overlay design program.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 28
Chapter-5
ANALYSIS AND RESULTS
5.1 DESIGNCASE DATABASE DEVELOPMENT
To build a design cases database, three different pavement cross sections have been
suggested: 3-layer, 4-layer, and 5-layer. For each run, the overlay design software “WINFLEX”
was used to determine an overlay thickness for the three cross sections based on controlling both
fatigue failure at the bottom of the bituminous layers and/or Rutting failure on the top of the
subgrade layer. The data range has already been suggested in 3.1. The data set was built taking
into consideration of temperature, seasonal variation. 21 design-cases were designed with 600
runs per database were done. Most of the parameters are tried to be varied as ANN doesn’t take
constant variables into account. Hence to check for all input variables varied data for each is
taken. Example of a output text file of WINFLEX is given at Appendix-A.
Figure 5-1 WINFLEX Program (WINFLEX, 2006)
5.2 TRAINING AND TESTING PROCESS FOR ACCURACY OF INPUT DATA
Here requirement is to get best mapping for the input to the desired output (Hov). Each
input set produces output under ANN with varied random set of initial weights. By training the
network, the weights of the system continually adjust to incrementally reduce the difference
between the output and the desired response. This difference is referred to as the error and here
measured as the Mean Squared Error (MSE). The MSE is the average of the squares of the
difference between each output and the desired output.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 29
An example of trained ANN dataset along with performance, training set and regression
plot can be seen at Appendix A.
5.3 SENSITIVITY ANALYSIS
The analysis indicates that changing the transfer function has a noticeable effect on the
accuracy. Furthermore, the number of hidden nodes has an effect on the accuracy, where using
more number of hidden nodes gives high accuracy. To achieve high accuracy, the number of
hidden nodes is preferable to be more than 25 nodes. On the other hand, ANN predicts much
better with the two hidden layers.
Hence for different transfer function (TRAINLM and TRAINSCG) for each design-cases
ANN analysis was done using MATLAB. The no of hidden nodes were also changed for each
transfer function. As number of nodes above 20 gives better results hence when plotted the
Percent Accuracy Vs no. of hidden nodes, the transfer function which gave better accuracy
between 20 to 50 was selected for that specific design-case. Graph between Percent accuracy and
no of hidden nodes are presented here for different cases where x_y means the database refers to
(x+1) layered initial structure and y design base, criteria of which is given earlier. For example,
1_2 means 3 layered initial structure and design case no. 2 which has been discussed in Section
4.1.
Figure5-2 Graph of Percent Accuracy Vs No. of Hidden
Nodes for 3layered pavement: Design-case 2
Figure 5-3 Graph of Percent Accuracy Vs No. of Hidden Nodes for 3layered Pavement: Design-case 3
10 20 50
TRAINLM 99.980897 99.988643 99.9912597
TRAINSCG 99.918669 99.72984 99.935667
99.5
99.6
99.7
99.8
99.9
100
100.1
Pe
rce
nt
Acc
ura
cy
1_2:Percent Accuracy Vs No. of Hidden Nodes
10 20 50
TRAINLM 99.987666 99.9967187 99.978117
TRAINSCG 99.965591 99.960703 99.978674
99.94
99.95
99.96
99.97
99.98
99.99
100
Pe
rce
nt
Acc
ura
cy
1_3:Percent Accuracy Vs No. of Hidden Nodes
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 30
Figure 5-4 Graph of Percent Accuracy Vs No. of Hidden Nodes for 3layered Pavement: Design-case 4
Figure 5-5 Graph of Percent Accuracy Vs No. of Hidden Nodes for 3layered Pavement: Design-case 5
Figure 5-6 Graph of Percent Accuracy Vs No. of Hidden Nodes for 3layered Pavement: Design-case 6
Figure 5-7 Graph of Percent Accuracy Vs No. of Hidden Nodes for 3layered Pavement: Design-case 7
Figure 5-8 Graph of Percent Accuracy Vs No. of Hidden Nodes for 4layered Pavement: Design-case 2
Figure 5-9 Graph of Percent Accuracy Vs No. of Hidden Nodes for 4layered Pavement: Design-case 3
10 20 50
TRAINLM 99.9957436 99.9947884 99.9946075
TRAINSCG 99.975381 99.94032 99.940339
99.9
99.92
99.94
99.96
99.98
100
Pe
rce
nta
ge A
ccu
racy
1_4: Percent Accuracy Vs No of
Hidden Nodes
10 20 50
TRAINLM 99.9972118 99.9926831 99.9930688
TRAINSCG 99.97054 99.960737 99.78332
99.6599.7
99.7599.8
99.8599.9
99.95100
100.05
Pe
rce
nt
Acc
ura
cy
1_5:Percent Accuracy Vs No. of Hidden Nodes
10 20 50
TRAINLM 99.951777 99.951374 99.944785
TRAINSCG 99.933503 99.906813 99.8282
99.75
99.8
99.85
99.9
99.95
100
Pe
rce
nt
Acc
ura
cy
1_6:Percent Accuracy Vs No. of Hidden Nodes
10 20 50
TRAINLM 99.9959637 99.9934139 99.9949656
TRAINSCG 99.956886 99.961401 99.982354
99.92
99.94
99.96
99.98
100
Pe
rce
nt
Acc
ura
cy
1_7:Percent Accuracy Vs No. of Hidden Nodes
10 20 30
TRAINLM 99.9952182 99.985609 99.956855
TRAINSCG 99.7525 99.929994 99.80458
99.6
99.7
99.8
99.9
100
100.1
Pe
rce
nt
Acc
ura
cy
2_2:Percent Accuracy Vs No. of Hidden Nodes
10 20 30
TRAINLM 99.9907114 99.956529 99.953329
TRAINSCG 99.6013 99.8205 99.914209
99.499.599.699.799.899.9100
100.1
Pe
rce
nt
Acc
ura
cy
2_3:Percent Accuracy Vs No. of Hidden Nodes
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 31
Figure 5-10 Graph of Percent Accuracy Vs No. of Hidden
Nodes for 4layered Pavement: Design-case 4
Figure 5-11 Graph of Percent Accuracy Vs No. of Hidden Nodes for 4layered Pavement: Design-case 5
Figure 5-12 Graph of Percent Accuracy Vs No. of Hidden Nodes for 4layered Pavement: Design-case 6
Figure 5-13 Graph of Percent Accuracy Vs No. of Hidden
Nodes for 4layered Pavement: Design-case 7
Figure 5-14 Graph of Percent Accuracy Vs No. of Hidden Nodes for 5layered Pavement: Design-case 2
Figure 5-15 Graph of Percent Accuracy Vs No. of Hidden
Nodes for 5layered Pavement: Design-case 3
10 20 50
TRAINLM 99.962875 99.96637 99.958478
TRAINSCG 99.933033 99.909476 99.936124
99.88
99.9
99.92
99.94
99.96
99.98
Pe
rce
nt
Acc
ura
cy
2_4:Percent Accuracy Vs No. of Hidden Nodes
10 20 50
TRAINLM 99.985806 99.980421 99.977429
TRAINSCG 99.97184 99.944215 99.950481
99.92
99.94
99.96
99.98
100
2_5:Percent Accuracy Vs No. of Hidden Nodes
10 20 50
TRAINLM 99.981727 99.979235 99.97007
TRAINSCG 99.982971 99.944041 99.967341
99.9299.9399.9499.9599.9699.9799.9899.99
Pe
rce
nt
Acc
ura
cy
2_6:Percent Accuracy Vs No. of Hidden Nodes
10 20 50
TRAINLM 99.9977148 99.9967648 99.9939235
TRAINSCG 99.988625 99.980263 99.8915
99.8
99.85
99.9
99.95
100
100.05
Pe
rce
nt
Acc
ura
cy
2_7:Percent Accuracy Vs No. of Hidden Nodes
10 20 30
TRAINLM 99.9952182 99.985609 99.956855
TRAINSCG 99.7525 99.929994 99.80458
99.6
99.7
99.8
99.9
100
100.1
Pe
rce
nt
Acc
ura
cy
3_2:Percent Accuracy Vs No. of Hidden Nodes
10 20 50
TRAINLM 99.987666 99.9967187 99.978117
TRAINSCG 99.965591 99.960703 99.978674
99.9499.9599.9699.9799.9899.99
100
Pe
rce
nt
Acc
ura
cy
3_3:Percent Accuracy Vs No. of Hidden Nodes
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 32
Figure 5-16 Graph of Percent Accuracy Vs No. of Hidden
Nodes for 5layered Pavement: Design-case 4
Figure 5-17 Graph of Percent Accuracy Vs No. of Hidden
Nodes for 5layered Pavement: Design-case 5
Figure 5-18 Graph of Percent Accuracy Vs No. of Hidden
Nodes for 5layered Pavement: Design-case 6
Figure 5-19 Graph of Percent Accuracy Vs No. of Hidden
Nodes for 5layered Pavement: Design-case 7
Hence finally for testing process, transfer function for multilayer perception to be considered is given
below. 3-LAYER DESIGN
CRITERIA
TRANSFER FUNCTION
1. Considering fatigue failure
in old pavement.
TRAINLM
2. Considering fatigue failure
in new overlay.
TRAINLM
3. Considering fatigue failure
in new overlay and old
pavement.
TRAINSCG
4. Considering both rutting
and fatigue failure in old
pavement.
TRAINLM
5. Considering both rutting
and fatigue failure in new
overlay.
TRAINSCG
10 20 50
TRAINLM 99.962875 99.96637 99.958478
TRAINSCG 99.933033 99.909476 99.936124
99.88
99.9
99.92
99.94
99.96
99.98
Pe
rce
nt
Acc
ura
cy
3_4:Percent Accuracy Vs No. of Hidden Nodes
10 20 50
TRAINLM 99.9972118 99.9926831 99.9930688
TRAINSCG 99.97054 99.960737 99.78332
99.6599.7
99.7599.8
99.8599.9
99.95100
100.05
Pe
rce
nt
Acc
ura
cy
3_5:Percent Accuracy Vs No. of Hidden Nodes
10 20 50
TRAINLM 99.981727 99.979235 99.97007
TRAINSCG 99.982971 99.944041 99.967341
99.9299.9399.9499.9599.9699.9799.9899.99
Pe
rce
nt
Acc
ura
cy
3_6:Percent Accuracy Vs No. of Hidden Nodes
10 20 50
TRAINLM 99.9959637 99.9934139 99.9949656
TRAINSCG 99.956886 99.961401 99.982354
99.9399.9499.9599.9699.9799.9899.99
100
Pe
rce
nt
Acc
ura
cy
3_7:Percent Accuracy Vs No. of Hidden Nodes
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 33
6. Considering both rutting
and fatigue failure in new
overlay and old pavement.
TRAINSCG
7. Considering rutting on the
sub grade layer.
TRAINLM
Table 5-1 Transfer Function for 3layer pavement design
4-LAYER DESIGN
CRITERIA
TRANSFER FUNCTION
1. Considering fatigue failure
in old pavement.
TRAINLM
2. Considering fatigue failure
in new overlay.
TRAINLM
3. Considering fatigue failure
in new overlay and old
pavement.
TRAINSCG
4. Considering both rutting
and fatigue failure in old
pavement.
TRAINSCG
5. Considering both rutting
and fatigue failure in new
overlay.
TRAINLM
6. Considering both rutting
and fatigue failure in new
overlay and old pavement.
TRAINSCG
7. Considering rutting on the
sub grade layer.
TRAINLM
Table 5-2 Transfer function for 4 layer pavement Design
5-LAYER DESIGN
CRITERIA
TRANSFER FUNCTION
1. Considering fatigue failure
in old pavement.
TRAINLM
2. Considering fatigue failure
in new overlay.
TRAINLM
3. Considering fatigue failure
in new overlay and old
pavement.
TRAINSCG
4. Considering both rutting
and fatigue failure in old
pavement.
TRAINLM
5. Considering both rutting
and fatigue failure in new
overlay.
TRAINLM
6. Considering both rutting
and fatigue failure in new
overlay and old pavement.
TRAINSCG
7. Considering rutting on the
sub grade layer.
TRAINLM
Table 5-3 Transfer function for 5 layer pavement Design
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 34
Another analysis was done to compare neural models: MLP networks versus Radial Basis
Function (RBF) networks. Results showed, MLP are more accurate than RBF networks The
reason being equal importance given to all input variables in RBF networks, which is not the
case with MLP networks. As such Weighting process of input variables is very much important
in design of flexible pavements. Mostly between 200 to 300 no of neurons, the RBF function
reaches required performance condition. This can be checked from examples of performance plot
given in Appendix B.
5.4 TESTING PROCESS ON ACTUAL DATA OBTAINED FROM MORTH R-81
Once the network has been trained, the trained network should be exposed to the data sets
obtained from ministry of Road Transport and Highways (Research scheme R-81) which gives
falling weight deflectometer test result of pavements tested by IIT kharagpur. Therefore, design
cases have been selected to represent training data sets distributed on the cross sections.
Comparison is to be done between predicted overlay thickness using the trained network and the
actual ones from MORTH R-81 to compute accuracy rate or reliability. If it is low, then the
network is not accurately taught and other training sets are required for retraining purpose,
otherwise, the network is considered consistent and ready for implementation.
In-service pavement sections
For the present study, some pavement sections in the states of West Bengal, Orissa and
Jharkhand were selected for detailed investigation. Specification as per Research Scheme R-81,
average daily two-way traffic on these roads ranged from 300 to 7000 commercial vehicles per
day (cvpd). The granular sub-base and base of in-service pavements consisted of layers of sand,
brickbat and crushed stone aggregates in varying thickness and they were treated as a single
layer (granular base) for analysis. Similarly, the bituminous surfacing layer consisted mostly of
bituminous macadam covered with premix carpet and seal coat. One or more layers of
bituminous material placed over the granular layer at different times with varied thicknesses
were also considered as one layer. Details of the selected test sections are given in Table.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 35
Table 5-4 Details of selected test sections (MORTH R-81)
The annual average rainfall in the region is about 1250mm and the pavement temperatures vary
in the range from 20oC to 50
oC. All the stretches have single carriageway carrying two-way
Traffic. The average shoulder width was in the range of 1 to 2 m. It was observed at the time of
investigation that some of the pavement sections were badly cracked and some were showing
cracks covering nearly 20% of the pavement area. (MORTH R-81)
Back-calculated layer moduli
Deflection data obtained using the FWD was used to back-calculate effective pavement
layer moduli and BACKGA program was used to compute the layer moduli. The pavement
sections were considered as three layer elastic systems consisting of bituminous surfacing,
granular base and subgrade. The inputs required for back-calculation analysis are the thicknesses
of the first two layers and Poisson ratio values of all the three layers. Thicknesses measured by
excavating test pits were used in the analysis. Since the moduli of granular bases and sub-bases
are not much different, two layers were considered as a single granular layer termed as Granular
Base (GB). Similarly, the thicknesses of different layers of bituminous materials were added for
getting the surface course thickness. Poisson ratio values of bituminous layer, granular layer and
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 36
subgrade were taken as 0.5, 0.4 and 0.4 respectively. The moduli ranges considered in the back-
calculation for different situations are given in Table.
Table 5-5 Modulus Range (MPa) for types of surfacing (MORTH R-81)
The surface loading considered for analysis is 40kN acting over a circular contract area
with a radius of 50mm. Surface deflections measured at radial distances of 0, 300, 600, 900,
1200, 1500 and 1800 mm were the main inputs to BACKGA. These deflections were normalized
to correspond to a load of 40 kN. The following GA parameters were used for the analysis.
Population Size = 60;
Maximum number of Generations = 60;
Probability of Crossover =0.74;
Probability of Mutation = 0.1 [MORTH R-81]
5.4.1 THIN OLD PAVEMENTS
Pavement sections with thickness of bituminous surfacing less than 75 mm were
considered as thin pavements (Thin PC) in this investigation. Back-calculated pavement layer
moduli, Sectional details, Surface deflections measured in different seasons using FWD are
specified for each stretch in Appendix C. Here provided is the ANN analysis of the same using it
as SAMPLE and previously obtained design database as INPUT.
When used with WINFLEX this gave the required overlay thickness which was checked with
predicted value using ANN and error is calculated. Given below are some of the selected cases
that were considered to check ANN’s predicting capacity. There are cases where Accuracy is
found to be more than 100%, it’s because during actual use the thickness of overlay taken is less
than code found thickness. When checked with ANN result these give greater than 100%
accuracy. From sensitivity analysis,
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 37
Design-case: “ Both rutting failure in subgrade and fatigue failure in new overlay or old
asphalt“
Transfer function: TRAINSCG.
CASE Thin_1: for Km 1.865 to 2.000 of SH * (Salua Road) for the Deflection Data Collected during the Year 2001-02. Location Km Temp (C) Eov(MPa) Eold(MPa) Eb(MPa) Esg(MPa) Hold(mm) Hb(mm) ESAL Hov(mm) Hov PREDICTED(mm) Accuracy(%)
2.000L 31 312.93 278.4 292.8 42.9 40 245 5000000 530.1247 473.3645378 89.29305
1.970L 31 312.93 302.1 288.8 42.3 45 264 5000000 516.9539 455.2845487 88.07063
1.940L 31 312.93 253.7 252.7 40.7 42 295 5000000 536.7101 450.6575404 83.96666
1.910L 31 312.93 248.8 227.9 40.7 43 253 5000000 559.759 467.1644197 83.45813
1.880L 31 312.93 276.3 266.4 46.2 52 279 5000000 513.6612 462.6918152 90.07724
1.865R 31 312.93 333.3 317.8 44.5 45 300 5000000 487.3196 443.2281918 90.95226
1.985R 31 312.93 336.5 256.3 43.1 46 270 5000000 523.5393 455.0231833 86.9129
1.955R 31 312.93 230.1 225.5 38.6 45 278 5000000 559.759 456.7093966 81.59036
1.925R 31 312.93 236.5 214.9 38.1 45 290 5000000 563.0517 455.2394029 80.85215
1.895R 31 312.93 320.4 269.5 39.7 52 300 5000000 510.3685 452.9436294 88.74835
2.000L 25 312.93 320.4 311.2 51 40 245 5000000 526.832 445.9600203 84.64938
1.970L 25 312.93 360.2 269.9 50.7 45 264 5000000 536.7101 448.4813899 83.5612
1.940L 25 312.93 346.2 309.9 48.2 42 295 5000000 513.6612 455.3030283 88.63878
1.910L 25 312.93 247.3 249.7 41.1 43 253 5000000 566.3444 474.372835 83.76049
1.880L 25 312.93 335.4 326.8 47.9 52 279 5000000 503.7831 450.6184391 89.44691
1.865R 25 312.93 341.3 286.6 44.1 45 300 5000000 526.832 455.6453741 86.48779
1.985R 25 312.93 321.6 291 43 46 270 5000000 533.4174 461.9540285 86.60273
1.955R 25 312.93 296.7 259.8 40 45 278 5000000 553.1736 464.1701211 83.91039
1.925R 25 312.93 283.8 306 42.7 45 290 5000000 523.5393 461.4700107 88.14429
1.895R 25 312.93 254.8 316.5 47.5 52 300 5000000 503.7831 452.9421795 89.90817
2.000L 41 312.93 332.8 316.1 53 40 245 5000000 424.7583 409.6491858 96.44289
1.910L 41 312.93 270.9 283.5 42 43 253 5000000 454.3926 424.4681603 93.41441
1.955R 41 312.93 367.7 262 45.1 45 278 5000000 428.051 417.0534505 97.43079
5-6 CASE Thin_1: ANN data for Km 1.865 to 2.000 of SH * (Salua Road) for the Deflection Data Collected during the Year 2001-02.
Avg Accuracy: 91.428%
CASE Thin_2: for Km 2.850 to 3.000 of SH* (Salua Road) for the Deflection Data Collected during the Year 2001-02. Location Km Temp
(C) Eov (MPa)
Eold (MPa)
Eb (MPa)
Esg (MPa)
Hold (mm.)
Hb (mm)
ESAL Hov (mm)
Hov PREDICTED (mm)
Accuracy (%)
3.000L 31 329.87 227.9 134.4 37.2 45 290 5000000 612.4422 598.8803706 97.78561
2.940L 31 329.87 241 197.7 40.8 45 295 5000000 566.3444 566.0938247 99.95576
2.910L 31 329.87 327.9 205.7 44.8 45 290 5000000 546.5882 541.283995 99.02958
2.940L 25 329.87 404.3 226.3 44.6 45 295 5000000 559.759 544.9175571 97.3486
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 38
2.910L 25 329.87 326.8 254.9 50.1 45 290 5000000 540.0028 525.4833337 97.31122
2.880L 25 329.87 289.2 278.3 53.8 45 270 5000000 530.1247 517.8199402 97.67889
2.850L 25 329.87 455.9 243.5 43.3 45 280 5000000 549.8809 549.3203667 99.89806
2.985R 25 329.87 308.6 232.5 55.5 45 282 5000000 549.8809 508.625111 92.49732
2.955R 25 329.87 405.3 271.2 51.8 45 270 5000000 530.1247 514.5074945 97.05405
2.895R 25 329.87 387 231.2 43.6 45 290 5000000 559.759 553.1407086 98.81765
3.000L 41 329.87 296.7 231.3 44.8 45 290 5000000 451.0999 447.7585232 99.25928
2.970L 41 329.87 333.3 276.1 44.9 45 279 5000000 431.3437 427.8491979 99.18986
2.940L 41 329.87 384.9 284.4 54 45 295 5000000 405.0021 404.9861363 99.99606
2.880L 41 329.87 375.2 259.8 52.8 45 270 5000000 421.4656 420.9473756 99.87704
2.850L 41 329.87 461.2 298.9 50.2 45 280 5000000 398.4167 397.3088509 99.72194
2.985R 41 329.87 338.7 308.7 57.4 45 282 5000000 405.0021 402.7085277 99.43369
2.925R 41 329.87 301 231.2 51.6 45 280 5000000 447.8072 439.5353365 98.15281
2.895R 41 329.87 395.4 245.8 52 45 290 5000000 421.4656 419.7968146 99.60405
5-7 CASE Thin_2: ANN data for Km 2.850 to 3.000 of SH* (Salua Road) for the Deflection Data Collected during the Year 2001-02.
Avg.Accuracy:99.583%
Case Thin_3: for Km 4.625 to 5.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-02. Location Km Temp (C) Eov
(MPa) Eold
(MPa) Eb
(MPa) Esg
(MPa) Hold (mm)
Hb (mm)
ESAL Hov (mm)
Hov PREDICTED (mm)
Accuracy (%)
5L 31 308.47 246.2 189 39 65 300 5000000 553.1736 501.8115004 90.71501
4.970L 31 308.47 216.1 191.6 41.9 65 250 5000000 563.0517 502.4126166 89.23028
4.940L 31 308.47 287 151.6 44.1 65 300 5000000 559.759 507.1280092 90.59756
4.910L 31 308.47 239.7 180.2 48.2 60 390 5000000 536.7101 504.6410464 94.02488
4.695L 31 308.47 265.5 182.4 50.8 63 360 5000000 530.1247 490.6132866 92.54677
4.925R 31 308.47 297.8 163.4 41.7 67 395 5000000 540.0028 498.725505 92.3561
4.9R 31 308.47 264.3 184.9 42.3 60 320 5000000 549.8809 507.5506819 92.30193
4.705R 31 308.47 233.3 193.4 40.8 60 345 5000000 546.5882 509.4084729 93.19785
4.645R 31 308.47 261.2 173.1 40.4 65 395 5000000 543.2955 504.5831208 92.87453
4.625R 31 308.47 231.1 157.7 39.6 65 350 5000000 566.3444 513.6313554 90.6924
5L 25 308.47 262.3 194.7 41.8 65 300 5000000 572.9298 500.1308245 87.29356
4.970L 25 308.47 290.3 240.9 54.6 65 250 5000000 540.0028 479.1519223 88.73138
4.940L 25 308.47 337.6 214.5 55.6 65 300 5000000 540.0028 478.790504 88.66445
4.910L 25 308.47 387 190.7 58.8 60 390 5000000 540.0028 491.098148 90.94363
4.695L 25 308.47 329 214.9 61.1 63 360 5000000 526.832 478.3732172 90.80185
4.925R 25 308.47 387.2 192.5 53.6 67 395 5000000 536.7101 482.5067326 89.90081
4.9R 25 308.47 290.3 235.1 54 60 320 5000000 533.4174 486.0155778 91.11356
4.705R 25 308.47 241.9 215.3 49.2 60 345 5000000 546.5882 505.4560198 92.47474
4.645R 25 308.47 234.4 209.6 50.2 65 395 5000000 536.7101 499.5972026 93.08511
4.625R 25 308.47 305.3 252.7 53.5 65 350 5000000 513.6612 471.8363183 91.8575
5L 41 308.47 373.1 227.2 45 65 300 5000000 401.7094 396.7964893 98.777
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 39
4.970L 41 308.47 298.9 236 55.1 65 250 5000000 411.5875 401.4183113 97.52928
4.940L 41 308.47 395.6 203 58.1 65 300 5000000 395.124 386.6609507 97.85813
4.910L 41 308.47 411.8 189 60.7 60 390 5000000 395.124 383.5635296 97.07422
4.695L 41 308.47 368.8 202.2 66.5 63 360 5000000 391.8313 372.3786021 95.03544
4.925R 41 308.47 391.3 186.8 55.7 67 395 5000000 391.8313 378.6495023 96.63585
4.9R 41 308.47 392.4 236.5 54.5 60 320 5000000 391.8313 381.6827018 97.40996
4.705R 41 308.47 286.6 206.1 50.7 60 345 5000000 421.4656 407.2239297 96.62092
4.645R 41 308.47 370.9 219.7 57.2 65 395 5000000 381.9532 372.2135667 97.45005
4.625R 41 308.47 358 232.9 57.7 65 350 5000000 385.2459 369.9210455 96.02206
5-8 Case Thin_3:ANN data for Km 4.625 to 5.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-02.
Avg.Accuracy:93.127%
CASE Thin_4: for Km 3.370 to 4.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-02
Location Km Temp (C) Eov (MPa)
Eold (MPa)
Eb( MPa)
Esg (MPa)
Hold (mm)
Hb (mm)
ESAL Hov (mm)
Hov PREDICTED (mm)
Accuracy (%)
4.000L 31 464.51 406.7 155.9 42.5 68 285 5000000 530.1247 496.2128286 93.60304
3.985R 31 464.51 330.8 144.3 43.6 67 282 5000000 546.5882 506.0797607 92.58886
3.970L 31 464.51 506.3 130.1 36.9 68 480 5000000 516.9539 488.3848731 94.47358
3.955R 31 464.51 430.6 103.7 35.7 61 390 5000000 559.759 524.2414069 93.65484
3.940L 31 464.51 380.1 101.7 37.5 60 390 5000000 566.3444 531.4542801 93.83942
3.925R 31 464.51 460.3 100.9 43.3 65 370 5000000 546.5882 503.8281945 92.17692
3.910L 31 464.51 439.1 139.3 41.3 67 280 5000000 536.7101 502.3148555 93.59147
3.710R 31 464.51 390.3 124.4 43.7 66 280 5000000 549.8809 510.601368 92.85672
3.695L 31 464.51 400.1 130.7 49.1 70 400 5000000 523.5393 481.7938832 92.02631
3.680R 31 464.51 444.2 100.7 46.9 67 390 5000000 540.0028 497.8678695 92.19728
3.665L 31 464.51 380.7 131.2 44 64 280 5000000 546.5882 511.3904677 93.56047
3.650R 31 464.51 360.7 126.5 44.4 68 290 5000000 546.5882 507.9300259 92.92737
3.400L 31 464.51 340.2 114.1 68.4 70 430 5000000 526.832 465.5351464 88.36501
3.385R 31 464.51 330.5 115.5 61.9 70 420 5000000 530.1247 475.9308149 89.77714
3.370L 31 464.51 490.7 116.3 72 72 420 5000000 497.1977 437.3461034 87.96221
4.000L 25 464.51 412.5 165.7 52.7 68 285 5000000 553.1736 492.5401053 89.03898
3.985R 25 464.51 400.8 146.7 66.6 67 282 5000000 553.1736 486.4891655 87.94512
3.970L 25 464.51 453.3 130.7 41 68 480 5000000 566.3444 520.8481981 91.96669
3.955R 25 464.51 454.7 119.7 41.3 61 390 5000000 586.1006 527.0051136 89.91718
3.940L 25 464.51 438.3 159.2 43 60 390 5000000 559.759 516.9116847 92.3454
3.925R 25 464.51 360.6 121.4 40.3 65 370 5000000 592.686 524.9581656 88.57273
3.910L 25 464.51 599.4 120.1 56 67 280 5000000 563.0517 492.257148 87.42663
3.710R 25 464.51 607 156.7 49 66 280 5000000 549.8809 489.7148312 89.05835
3.695L 25 464.51 505.2 177.3 66 70 400 5000000 513.6612 454.5319284 88.48866
3.680R 25 464.51 451.6 206.4 52.9 67 390 5000000 513.6612 472.4640718 91.97971
3.665L 25 464.51 533.3 146.1 46.1 64 280 5000000 566.3444 500.5779422 88.38755
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 40
3.650R 25 464.51 547.5 125.4 63.2 68 290 5000000 556.4663 485.3492042 87.21987
3.400L 25 464.51 643.1 269.5 73.5 70 430 5000000 447.8072 398.5322169 88.99638
3.385R 25 464.51 341.2 215.4 75 70 420 5000000 493.905 443.0285184 89.69914
3.370L 25 464.51 560 274.1 80 72 420 5000000 444.5145 391.8814265 88.15942
4.000L 41 464.51 412 189.8 61.7 68 285 5000000 395.124 387.1366956 97.97853
3.985R 41 464.51 404.4 161.1 69.7 67 282 5000000 401.7094 389.2200996 96.89096
3.970L 41 464.51 488 120.2 56.5 68 480 5000000 401.7094 391.655995 97.49734
3.955R 41 464.51 460.1 119 45.8 61 390 5000000 431.3437 417.0680278 96.69042
3.940L 41 464.51 549.5 145.1 53 60 390 5000000 401.7094 389.0892107 96.85838
3.925R 41 464.51 368.4 117.9 43 65 370 5000000 444.5145 426.8814748 96.03319
3.910L 41 464.51 501.1 170.2 65.5 67 280 5000000 388.5386 385.5842253 99.23962
3.710R 41 464.51 477.1 150.3 56 66 280 5000000 408.2948 406.3044146 99.51251
3.695L 41 464.51 594 189.2 77.3 70 400 5000000 352.3189 330.1705548 93.71355
3.680R 41 464.51 464.7 179.5 65.9 67 390 5000000 378.6605 361.8807867 95.56867
3.665L 41 464.51 454.6 168.9 48.7 64 280 5000000 414.8802 409.9455999 98.8106
3.650R 41 464.51 646.2 161.5 72.4 68 290 5000000 365.4897 362.5524529 99.19635
3.400L 41 464.51 627.1 271.3 77.1 70 430 5000000 316.0992 309.2812492 97.8431
3.385R 41 464.51 635 272.1 65.2 70 420 5000000 322.6846 315.1739329 97.67244
3.370L 41 464.51 649.5 269.4 88.5 72 420 5000000 309.5138 298.7257169 96.51451
5-9 CASE Thin_4: ANN data for Km 3.370 to 4.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-02
Avg.Accuracy:92.9516%
CASE Thin_5: for Km 15.000 to 15.270 of NH-60 for the Deflection Data Collected during the Year 2001-02
Location
Km
Temp
(C)
Eov
(MPa)
Eold
(MPa)
Eb
(MPa)
Esg
(MPa)
Hold
(mm)
Hb
(mm)
ESAL Hov
(mm)
Hov PREDICTED
(mm)
Accuracy
(%)
15.090R 31 321.27 302 172.9 49.5 61 373 5000000 533.4174 521.479065 97.76191
15.120L 31 321.27 298 145.8 40.9 50 373 5000000 572.9298 572.7875235 99.97517
15.150R 31 321.27 357.1 146 39.7 49 311 5000000 572.9298 567.3541474 99.02682
15.180L 31 321.27 243.8 150.1 35.1 42 291 5000000 605.8568 604.1041327 99.71071
15.240L 31 321.27 328.3 141.9 39.5 35 340 5000000 592.686 579.6577884 97.80184
15.270R 31 321.27 268.4 139.4 40.1 36 318 5000000 602.5641 596.3714095 98.97228
15.090R 25 321.27 301.2 134.7 44.6 61 373 5000000 605.8568 599.9107949 99.01858
15.180L 25 321.27 485.9 145 42.6 42 291 5000000 609.1495 602.1567007 98.85204
15.240L 25 321.27 305 141.4 46.5 35 340 5000000 615.7349 471.1958493 76.52577
15.270R 25 321.27 411 156 41.1 36 318 5000000 609.1495 555.7219941 91.22916
15.030R 41 321.27 276.2 182.3 43.5 50 370 5000000 457.6853 453.4063378 99.06509
15.060L 41 321.27 216.2 289.4 46.3 50 375 5000000 414.8802 413.5405411 99.6771
15.090R 41 321.27 282.1 189.2 45.4 61 373 5000000 431.3437 420.5125406 97.48897
15.120L 41 321.27 203.1 129.3 50.6 50 373 5000000 493.905 484.8574459 98.16816
15.150R 41 321.27 228.5 167.5 46.6 49 311 5000000 480.7342 468.8628235 97.53057
15.180L 41 321.27 234.8 169.2 47.1 42 291 5000000 493.905 486.9194667 98.58565
15.210R 41 321.27 321.2 153.9 48.5 41 334 5000000 480.7342 478.4751486 99.53008
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 41
15.240L 41 321.27 436.9 149.3 51.6 35 340 5000000 474.1488 470.282895 99.18466
15.270R 41 321.27 468.3 154.8 43.2 36 318 5000000 477.4415 472.393318 98.94266
5-10 CASE Thin_5: ANN data for Km 15.000 to 15.270 of NH-60 for the Deflection Data Collected during the Year 2001-02
Avg.Accuracy:98.898%
5.4.2 THICK OLD PAVEMENTS The pavement sections considered in the present study also included a number of thick
pavements bituminous layer thickness with more than 75 mm. The bituminous surface consisted
of repeated application of bituminous macadam (BM) and premix carpet over a period of time.
Sectional details of the pavements, Deflection data and backcalculated layer moduli are provided
at Appedix C. Thickness of bituminous layer thicknesses ranged from 90 to 200 mm. A few
sections had bituminous concrete overlay of 50 mm thickness. The thicknesses granular base
varied from 300 to 690 mm. Results from ANN analysis is given below for each Section.
When used with WINFLEX this gave the required overlay thickness which was checked
with predicted value using ANN and error is calculated.
From sensitivity analysis,
Design-case: “ only fatigue cracking in new overlay or old asphalt and rutting failure in
subgrade“
Transfer function: TRAINLM.
Since the thickness of the granular layer is very large (585 mm), Road Research R-56
[1999] has clearly established that only fatigue criteria will be applicable to such cases and
number of standard axle load for causing 20 mm rutting is very large.
CASE Thick_1:
for Km 123.795 to 124.000 of NH-6 for the Deflection Data collected during the Year 2000-01
(Kumar, 2001)
Location Km Temp
(C)
Eov
(MPa)
Eold
(MPa)
Eb(
MPa)
Esg
(MPa)
Hold
(mm)
Hb
(mm)
ESAL Hov
(mm)
Hov PREDICTED
(mm)
Accuracy
(%)
124.000L 31 761.97 721.6 209 55 95 134.9716 10000000 510.3685 495.7778296 97.14115
123.995R 31 761.97 742 160.7 49.5 95 130.925 10000000 523.5393 506.0729502 96.66379
123.950L 31 761.97 640.5 180.4 46.3 95 119.061 10000000 540.0028 500.6662731 92.7155
123.945R 31 761.97 841.9 164.4 41.6 95 145.9509 10000000 523.5393 503.6910726 96.20884
123.900L 31 761.97 719.4 159.4 36 95 127.4587 10000000 549.8809 498.4779016 90.65198
123.895R 31 761.97 802.2 162.8 39.6 95 139.9609 10000000 530.1247 499.1641357 94.15976
123.850L 31 761.97 854.4 200.2 41.9 95 152.9562 10000000 510.3685 495.5176829 97.09018
123.845R 31 761.97 866.5 137.5 36.7 95 145.6173 10000000 533.4174 507.0363565 95.05433
123.800L 31 761.97 500.4 210.5 32.1 95 103.1069 10000000 582.8079 467.8207394 80.27014
123.795R 31 761.97 761.6 170.4 35.5 95 135.1746 10000000 543.2955 493.7858636 90.88716
124.000L 25 761.97 956.2 242.9 56.1 95 173.9141 10000000 526.832 459.7753823 87.27173
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 42
123.995R 25 761.97 856.2 252.9 55.2 95 160.8607 10000000 536.7101 454.0580636 84.60025
123.950L 25 761.97 690.2 233.7 54.1 95 133.9998 10000000 563.0517 458.4101341 81.41528
123.945R 25 761.97 995.4 380.6 72.2 95 199.5711 10000000 464.2707 446.1210969 96.09073
123.900L 25 761.97 954.8 399.3 63.1 95 196.3948 10000000 470.8561 433.5689367 92.08099
123.895R 25 761.97 745.3 251.3 58.4 95 144.544 10000000 546.5882 460.5460368 84.25832
123.850L 25 761.97 790.4 260.5 63.2 95 152.4196 10000000 530.1247 460.0519328 86.78183
123.845R 25 761.97 968.5 396.5 80 95 197.9757 10000000 457.6853 441.4487754 96.45247
123.800L 25 761.97 893.9 393.1 51.2 95 186.6628 10000000 493.905 417.6567769 84.56217
123.795R 25 761.97 543.5 384.2 55 95 134.551 10000000 536.7101 425.8619884 79.34674
5-11 CASE Thick_1: ANN data for Km 123.795 to 124.000 of NH-6 for the Deflection Data collected during the Year 2000-01 (Kumar, 2001)
Avg.Accuracy:99.27%
CASE Thick_2: for Km 125.000 to 125.270 of NH-6 for the Deflection Data collected during Summer Season of the Year 2001-02. Location Km Temp
(C)
Eov
(MPa)
Eold
(MPa)
Eb
(MPa)
Esg
(MPa)
Hold
(mm)
Hb
(mm)
ESAL Hov
(mm)
Hov PREDICTED
(mm)
Accuracy
(%)
125.000L 31 478.7 603.7 268.5 66.4 100 614.9977 10000000 498.6617 428.051 85.83996
125.030R 31 478.7 595.9 247 62.4 100 619.9977 10000000 494.5416 441.2218 89.21834
125.060L 31 478.7 359.3 291.2 62.2 100 624.9977 10000000 478.7385 451.0999 94.22678
125.090R 31 478.7 488.4 296.3 63.8 100 619.9977 10000000 477.1178 428.051 89.716
125.120L 31 478.7 355.4 295.5 59.8 100 619.9977 10000000 478.8325 447.8072 93.52064
125.150R 31 478.7 618.5 276.7 63.2 100 639.9976 10000000 483.9381 421.4656 87.09082
125.180L 31 478.7 450.2 268.1 62.6 100 624.9977 10000000 489.7616 447.8072 91.4337
125.210R 31 478.7 532.3 230.6 59 100 624.9977 10000000 494.9887 460.978 93.12899
125.240L 31 478.7 375.9 278.7 60.3 100 629.9977 10000000 486.3162 454.3926 93.43562
125.270R 31 478.7 407.2 284.5 59.7 100 624.9977 10000000 481.0252 447.8072 93.09434
5-12 CASE Thick_2: ANN data for Km 125.000 to 125.270 of NH-6 for the Deflection Data collected during Summer Season of the Year 2001-02.
Avg.Accuracy:91.07%
CASE Thick_3: for Km 134.000 to 134.270 of NH-6 for the Deflection Data collected during the Year 2001-2002. Location Km Temp (C) Eov
(MPa)
Eold
(MPa)
Eb
(MPa)
Esg
(MPa)
Hold
(mm)
Hb
(mm)
ESAL Hov
(mm)
Hov PREDICTED
(mm)
Accuracy
(%)
134.000L 31 363.26 311.8 156.8 50.5 145 544.998 10000000 514.0629 516.9539 100.5624
134.030R 31 363.26 352.8 172.4 54.6 145 539.998 10000000 506.974 490.6123 96.77267
134.120L 31 363.26 329 148.9 53.5 145 559.9979 10000000 516.8361 513.6612 99.38571
134.180L 31 363.26 280.6 178 49 145 544.998 10000000 509.191 507.0758 99.58459
134.210R 31 363.26 372 169.2 35.3 145 534.998 10000000 480.969 510.3685 106.1126
134.240L 31 363.26 308.6 152.9 51.2 145 539.998 10000000 516.1497 520.2466 100.7937
134.000L 41 363.26 447.3 217.1 66.6 145 544.998 10000000 420.2466 322.6846 76.78458
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 43
134.030R 41 363.26 424.7 196 71.5 145 539.998 10000000 453.4039 332.5627 73.348
134.060L 41 363.26 402.1 230.7 60.6 145 544.998 10000000 395.8248 325.9773 82.35395
134.090R 41 363.26 348.3 204.4 53.6 145 537.998 10000000 388.4469 352.3189 90.69937
134.120L 41 363.26 353.7 197.3 60.1 145 559.9979 10000000 409.5879 349.0262 85.214
134.150R 41 363.26 304.6 209.8 62.3 145 557.9979 10000000 420.377 352.3189 83.81022
134.180L 41 363.26 376.3 265.1 64.2 145 544.998 10000000 400.1394 319.3919 79.82016
134.210R 41 363.26 365.5 215.8 68.4 145 534.998 10000000 436.2103 335.8554 76.99392
134.240L 41 363.26 396.7 250.1 64 145 539.998 10000000 402.6937 322.6846 80.13152
134.270R 41 363.26 364.5 223.7 65.1 145 549.998 10000000 418.1329 332.5627 79.53517
5-13 CASE Thick_3: ANN Data for Km 134.000 to 134.270 of NH-6 for the Deflection Data collected during the Year 2001-2002.
Avg.Accuracy:99.85%
CASE Thick_4: for Km 150.000 to 150.245 of NH-6 for the Deflection Data collected during the Year 2001-02 Location Km Temp (C) Eov
(MPa)
Eold
(MPa)
Eb
(MPa)
Esg
(MPa)
Hold
(mm)
Hb
(mm)
ESAL Hov
(mm)
Hov PREDICTED(mm) Accuracy
(%)
150.000L 31 366.87 254.8 203.9 44.9 170 705.9974 10000000 460.978 451.0436558 97.84494
150.060L 31 366.87 404.6 192.9 44.9 170 544.998 10000000 447.8072 446.0985014 99.61843
150.120L 31 366.87 384.9 162.1 48 170 534.998 10000000 474.1488 468.0078975 98.70486
150.240L 31 366.87 377.4 169.2 47.4 170 534.998 10000000 470.8561 466.2313766 99.01781
150.005R 31 366.87 375.2 162.1 47.6 170 534.998 10000000 474.1488 469.4604235 99.0112
150.065R 31 366.87 393.5 139.5 44.8 170 684.9975 10000000 480.7342 465.9809495 96.9311
150.245R 31 366.87 266.6 134.4 47.9 170 534.998 10000000 523.5393 478.1953157 91.33895
150.000L 25 366.87 401 198.6 73.1 170 705.9974 10000000 484.0269 371.8436037 76.82292
150.060L 25 366.87 375.2 239.1 57.4 170 544.998 10000000 474.1488 419.4227519 88.45804
150.120L 25 366.87 329 169.9 55.8 170 534.998 10000000 540.0028 399.8810961 74.05167
150.180L 25 366.87 375.2 190.7 54.9 170 574.9979 10000000 513.6612 416.7294829 81.12925
150.240L 25 366.87 448.3 210.9 56.9 170 534.998 10000000 487.3196 434.7305134 89.2085
150.005R 25 366.87 489.2 328.4 63.1 170 534.998 10000000 401.7094 374.0588003 93.11677
150.065R 25 366.87 375.2 322.7 57.4 170 684.9975 10000000 405.0021 389.4219748 96.15308
150.125R 25 366.87 376.3 218.4 63.6 173 664.9975 10000000 474.1488 382.6871245 80.71034
150.185R 25 366.87 423.6 274.7 70 170 534.998 10000000 437.9291 387.6045624 88.50852
150.245R 25 366.87 387.2 248.9 59.4 170 534.998 10000000 464.2707 417.4523487 89.91572
5-14 CASE Thick_4: ANN data for Km 150.000 to 150.245 of NH-6 for the Deflection Data collected during the Year 2001-02
Avg.Accuracy:98.25% CASE Thick_5: for Km 151.000 to 151.245 of NH-6 for the Deflection Data collected during the Year 2001-02
Location Km Temp
(C)
Eov
(MPa)
Eold
(MPa)
Eb
(MPa)
Esg
(MPa)
Hold
(mm)
Hb
(mm)
ESAL Hov
(mm)
Hov PREDICTED
(mm)
Accuracy
(%)
151.005R 31 400.13 501.3 201.6 46.6 170 469.9983 10000000 431.3437 410.3830613 95.14062
151.060L 31 400.13 306.1 264.6 32.6 170 545.998 10000000 434.6364 434.5238559 99.97411
151.065R 31 400.13 424 236 75.4 170 539.998 10000000 401.7094 391.4814664 97.4539
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 44
151.180L 31 400.13 361.2 189.5 48 170 489.9982 10000000 464.2707 462.7589019 99.67437
151.185R 31 400.13 408.2 265.3 47.2 173 474.9982 10000000 408.2948 404.7379867 99.12886
151.240L 31 400.13 324.7 202.6 74 170 534.998 10000000 441.2218 415.7991649 94.23813
151.000L 25 400.13 422.5 246.1 70 170 489.9982 10000000 460.978 400.290612 86.83508
151.005R 25 400.13 358 246.6 57.4 170 469.9983 10000000 477.4415 426.0573216 89.2376
151.060L 25 400.13 374.1 212.3 64.1 171 545.998 10000000 487.3196 402.8976887 82.67627
151.065R 25 400.13 375.2 204.3 57.4 170 539.998 10000000 500.4904 418.5411919 83.62622
151.120L 25 400.13 412.9 317 73.7 170 469.9983 10000000 418.1729 370.6786837 88.64245
151.125R 25 400.13 346.2 201.7 57.3 170 459.9983 10000000 513.6612 421.1626429 81.9923
151.180L 25 400.13 339.7 190.9 54.3 170 489.9982 10000000 523.5393 418.3471965 79.90751
151.185R 25 400.13 324.7 271.7 57.6 173 474.9982 10000000 460.978 416.7215611 90.39945
151.240L 25 400.13 340.8 218.9 63.6 170 534.998 10000000 490.6123 399.6192902 81.45317
151.245R 25 400.13 375.2 270.3 63.3 170 614.9977 10000000 441.2218 394.307613 89.36721
5-15 CASE Thick_5: ANN data for Km 151.000 to 151.245 of NH-6 for the Deflection Data collected during the Year 2001-02
Avg.Accuracy:97.73%
CASE Thick_7: for Km 152.000 to 152.245 of NH-6 for the Deflection Data collected during the Year 2001-02
Location
Km
Cross
section
Temp
(C)
Eov
(MPa)
Eold
(MPa)
Eb
(MPa)
Esg
(MPa)
Hold
(mm)
Hb
(mm)
ESAL Hov
(mm)
Hov
PREDICTED
(mm)
Accuracy
(%)
152.000L 3 31 397.83 372 199.1 47.8 170 546.99
8
10000000 455.3433
447.8072 98.34497
152.240L 3 31 397.83 410.7 181 59.1 170 544.99
8
10000000 459.1589
444.5145 96.81061
152.005R 3 31 397.83 308.6 201.7 49.4 170 543.99
8
10000000 467.2065
460.978 98.66685
152.065R 3 31 397.83 423.6 205.2 44.7 170 591.99
78
10000000 433.9455
431.3437 99.40043
152.125R 3 31 397.83 381.7 227.2 55.7 170 516.99
81
10000000 440.9743
424.7583 96.32268
152.185R 3 31 397.83 350.5 186.3 44.9 173 549.99
8
10000000 459.085
460.978 100.4123
152.245R 3 31 397.83 358 198.6 46 170 577.99
79
10000000 457.0038
451.0999 98.70812
152.000L 3 41 397.83 451.6 371.9 88.1 170 546.99
8
10000000 458.5471
237.0744 51.70121
152.060L 3 41 397.83 511.8 274.7 82.5 170 529.99
8
10000000 450.3846
263.416 58.48691
152.120L 3 41 397.83 405.3 201.3 57.4 170 627.99
77
10000000 355.6898
309.5138 87.0179
152.180L 3 41 397.83 382.7 203.5 58.6 170 529.99
8
10000000 363.1026
316.0992 87.05505
152.240L 3 41 397.83 478.4 240 83.5 170 544.99
8
10000000 454.293
279.8795 61.60771
152.005R 3 41 397.83 411.8 185.9 67.2 170 543.998 10000000 397.7194
312.8065 78.65006
152.065R 3 41 397.83 361.2 223.3 55 170 591.99
78
10000000 344.8566
312.8065 90.70626
152.125R 3 41 397.83 470.4 287.3 79.1 170 516.99
81
10000000 428.9684
266.7087 62.17443
152.185R 3 41 397.83 430.1 273.9 76 173 549.99
8
10000000 407.6098
273.2941 67.04796
152.245R 3 41 397.83 431.1 260.7 73 170 577.99
79
10000000 401.3656
279.8795 69.73182
5-16 CASE Thick_7: ANN data for Km 152.000 to 152.245 of NH-6 for the Deflection Data collected during the Year 2001-02
Avg.Accuracy:96.327%
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 45
CASE Thick_8: for Km 153.000 to 153.245 on NH-6 for the Deflection Data collected during the Year 2001-02 Location Km Temp (C) Eov
(MPa)
Eold
(MPa)
Eb
(MPa)
Esg
(MPa)
Hold
(mm)
Hb
(mm)
ESAL Hov
(mm)
Hov PREDICTED
(mm)
Accuracy
(%)
153.000L 31 505.47 814.5 204.8 57.5 170 557.9979 10000000 394.8484 368.7824 93.39848
153.060L 31 505.47 967.1 262 53.9 170 429.9984 10000000 385.895 342.4408 88.73937
153.240L 31 505.47 986.3 182.2 47.3 170 457.9983 10000000 405.8143 375.3678 92.49743
153.005R 31 505.47 543.5 163.8 37.9 170 574.9979 10000000 412.8309 444.5145 107.6747
153.065R 31 505.47 543.5 164.7 38.7 170 416.9984 10000000 413.8214 457.6853 110.5997
153.245R 31 505.47 388.3 167.8 45 170 444.9983 10000000 459.3886 477.4415 103.9298
153.000L 41 505.47 546.3 394.8 80 170 557.9979 10000000 461.6679 223.9036 48.49885
153.060L 41 505.47 672.3 398.3 79.9 170 429.9984 10000000 488.8166 220.6109 45.13163
153.120L 41 505.47 307.5 240.9 79.9 170 538.998 10000000 434.8789 309.5138 71.17241
153.180L 41 505.47 325.3 223.1 73.9 170 434.9984 10000000 419.2442 319.3919 76.18279
153.240L 41 505.47 317.1 334.6 79.6 170 457.9983 10000000 405.9187 283.1722 69.76082
153.005R 41 505.47 395.7 356.7 79.9 170 574.9979 10000000 418.4076 250.2452 59.80896
153.065R 41 505.47 388.3 205.5 79.6 170 416.9984 10000000 445.9874 312.8065 70.13798
153.125R 41 505.47 534.7 272.1 79.5 170 443.9983 10000000 433.9585 270.0014 62.21825
153.185R 41 505.47 453.3 244.3 79.9 173 464.9983 10000000 430.5286 286.4649 66.53795
153.245R 41 505.47 323.6 265.9 78 170 444.9983 10000000 419.8183 306.2211 72.94134
5-17 CASE Thick_8: ANN data for Km 153.000 to 153.245 on NH-6 for the Deflection Data collected during the Year 2001-02
Avg.Accuracy: 98.15%
5.4.3 EVALUATION OF COLD-MIX RECYCLED PAVEMENT SECTION One pavement section on NH-6 was considered for the structural evaluation before and
after cold-mix recycling was done using cement and bitumen emulsion. Using cold mix recycling
process, it is possible to rectify asphalt-aging problems and also to improve the structural
condition of the pavement. In the case of high traffic volume it is normally necessary to go for
overlay of one or more layers of hot-mixed asphalt on top of the recycled layer to further
strengthen the pavement to serve for the design life.
The pavement section Km 131.000 on NH-6 consisted of different layers of bituminous
material with a thickness of about 200 mm. Water Bound Macadam (WBM), morum, boulders
and sand formed the base course. The deflection data collected before recycling and after 7 and
28 days of recycling was used to compute the effective layer moduli. Table gives the moduli
values for the recycled section.
Layer Moduli of Recycled Pavement Stretch
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 46
Table 5-18 back calculated layer Moduli Range for Recycled Pavement Stretch (MORTH R-81)
The average modulus of the existing bituminous layer before recycling is about 390 Mpa
indicating low strength. This is due to the presence of excessive fatigue cracks on the surface of the
bituminous layer. The modulus value of the recycled layer has been found to increase to about 560 MPa.
Recycling resulted in significant increase in strength as evident even from the 7-day modulus value.
Location
Km
Time Temp
(C)
Eov
(MPa)
Eold
(MPa)
Eb
(MPa)
Esg
(MPa)
Hold
(mm)
Hb
(mm)
ESAL Hov
(mm)
Hov
PREDICTED
(mm)
Accurac
y
(%)
131.02 Just before 31 388.8 305.5 233.7 55.2 100 419.9
984
10000000 547.00961
97
520.2466 95.1074
131.04 Just before 31 388.8 323.9 241.2 60.4 100 419.9
984
10000000 542.43035
44
507.0758 93.48219
131.06 Just before 31 388.8 482 249.7 56.3 100 429.9
984
10000000 492.87142
92
480.7342 97.53745
131.08 Just before 31 388.8 304.8 235.7 56.4 100 434.9
984
10000000 546.62153
01
513.6612 93.97017
131.105 Just before 31 388.8 358.8 253.3 60.5 100 419.9
984
10000000 527.07866
28
493.905 93.70613
131.12 Just before 31 388.8 486.8 217.3 58.6 100 424.9
984
10000000 512.93437
57
493.905 96.2901
131.14 Just before 31 388.8 343.8 257.1 60.3 100 424.9
984
10000000 527.36803
8
493.905 93.65471
131.16 Just before 31 388.8 306.8 216.9 51.9 100 419.9
984
10000000 550.98655
97
530.1247 96.21373
131.18 Just before 31 388.8 368.4 237.6 54 100 419.9
984
10000000 523.65557
55
507.0758 96.83384
131.2 Just before 31 388.8 607.2 214.8 60 100 429.9
984
10000000 499.47416
62
477.4415 95.58883
131.02 7 days after 31 495.1 410.2 240.3 58.8 100 419.9
984
10000000 520.14488
89
474.1488 91.15706
131.04 7 days after 31 495.1 434.1 223.1 60.2 100 419.9
984
10000000 526.86536
89
480.7342 91.24422
131.06 7 days after 31 495.1 680.4 233.7 57.2 100 429.9
984
10000000 481.97057
05
444.5145 92.22856
131.08 7 days after 31 495.1 458.1 248.9 58.6 100 434.9
984
10000000 504.12052
16
464.2707 92.09518
131.105 7 days after 31 495.1 631.2 222.3 60.1 100 419.9
984
10000000 495.69195
94
454.3926 91.66834
131.12 7 days after 31 495.1 518.9 238.4 58.9 100 424.9
984
10000000 498.59446
53
460.978 92.4555
131.14 7 days after 31 495.1 456.7 244.3 59.1 100 424.9
984
10000000 508.02408
87
467.5634 92.03568
131.16 7 days after 31 495.1 538.1 245.1 56.9 100 419.9
984
10000000 487.28030
22
457.6853 93.92649
131.18 7 days after 31 495.1 366.3 221.6 54.9 100 419.9
984
10000000 535.72628
46
493.905 92.19354
131.2 7 days after 31 495.1 456.7 213.7 61.5 100 429.9
984
10000000 529.57424
06
480.7342 90.77749
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 47
131.02 28 days
after
31 557.8 481.8 254.2 55.9 100 419.9
984
10000000 489.96350
06
474.1488 96.77227
131.04 28 days
after
31 557.8 549.6 253.1 56.6 100 419.9
984
10000000 482.16069
63
467.5634 96.97252
131.06 28 days
after
31 557.8 807.6 228.7 55.7 100 429.9
984
10000000 480.66633
28
447.8072 93.16384
131.08 28 days
after
31 557.8 781.6 219.6 55.9 100 434.9
984
10000000 483.73005
23
454.3926 93.93516
131.105 28 days
after
31 557.8 472.6 236.9 55.8 100 419.9
984
10000000 499.22038
65
484.0269 96.95656
131.12 28 days
after
31 557.8 517.6 227.1 58.8 100 424.9
984
10000000 503.01108
82
480.7342 95.57129
131.14 28 days
after
31 557.8 494.3 239.2 60.4 100 424.9
984
10000000 506.76433
47
477.4415 94.21371
131.16 28 days
after
31 557.8 575.1 223.2 58.2 100 419.9
984
10000000 494.20903
89
477.4415 96.6072
131.18 28 days
after
31 557.8 489.5 234.1 59.6 100 419.9
984
10000000 507.59213
1
480.7342 94.70876
131.2 28 days
after
31 557.8 408.1 266.1 58.7 100 429.9
984
10000000 506.01578
52
477.4415 94.35308
5-19 ANN data for Cold-mix Recycled pavement Stretch
Avg.Accuracy:
Just before recycling:95.238%
7days after recycling:91.978%
28days after recycling:95.325%
5.4.4 EFFECT OF TEMPERATURE ON PAVEMENT LAYER MODULI Pavement layer moduli, especially that of bituminous layer varies with pavement temperature. In
order to study the effect on pavement temperature on the back-calculated moduli, FWD tests were
conducted on a few sections. The pavement temperature was recorded at a depth of 25 mm during testing
at each location. The deflections were measured at temperatures of 25, 30, 35 and 40◦C.
The back-calculated layer moduli values are shown in the Table.
Layer Moduli for the Deflection Data collected on KM 112. 000 to 112.540 of NH-6 at Different
Pavement Temperatures
table 5-20 Layer Moduli for the Deflection Data collected on KM 112. 000 to 112.540 of NH-6 at Different Pavement
Temperatures (MORTH R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 48
Location Km Temp (C) Eov
(MPa)
Eold
(MPa)
Eb
(MPa)
Esg
(MPa)
Hold
(mm)
Hb
(mm)
ESAL Hov
(mm)
Hov PREDICTED
(mm)
Accuracy
(%)
112 41 863.3 755.1 289.2 67.1 95 554.9979 10000000 329.27 234.291946 71.15496
112.06 41 863.3 885.4 282.9 68.9 95 542.998 10000000 319.3919 186.7475134 58.46971
112.12 41 863.3 775.1 313.1 66.4 95 549.998 10000000 319.3919 230.5773873 72.19262
112.18 41 863.3 778.6 313.1 65.6 95 539.998 10000000 322.6846 233.131469 72.24747
112.24 41 863.3 823.5 317.4 68.1 95 539.998 10000000 316.0992 210.9650918 66.74015
112.3 41 863.3 861.8 295.8 65.3 95 559.9979 10000000 319.3919 200.3087934 62.71568
112.36 41 863.3 1031.9 285.7 69.5 95 564.9979 10000000 306.2211 124.8815405 40.78149
112.42 41 863.3 898.2 295.5 67.5 95 409.9985 10000000 325.9773 199.5391869 61.2126
112.48 41 863.3 896.7 293.4 69.1 95 419.9984 10000000 325.9773 194.7504955 59.74358
112.54 41 863.3 926.4 292.4 66.1 95 519.9981 10000000 316.0992 181.9450528 57.55948
112 35 1041.6 1037.5 285.7 64.9 95 554.9979 10000000 355.6116 212.6968201 59.81155
112.06 35 1041.6 1016.4 303.3 66.8 95 542.998 10000000 352.3189 218.0637766 61.89386
112.12 35 1041.6 1038.7 287.3 65.1 95 549.998 10000000 355.6116 213.3037151 59.98222
112.18 35 1041.6 1049.8 281.4 63.1 95 539.998 10000000 358.9043 216.8542789 60.4212
112.24 35 1041.6 1003.5 297.9 63.8 95 539.998 10000000 355.6116 228.6896952 64.30884
112.3 35 1041.6 1031.2 301.7 66.4 95 559.9979 10000000 352.3189 210.1219163 59.63969
112.36 35 1041.6 1007 275.9 66.7 95 564.9979 10000000 362.197 215.4433847 59.48238
112.42 35 1041.6 937.2 259.1 68.9 95 409.9985 10000000 381.9532 246.258318 64.47343
112.48 35 1041.6 1221.8 235.7 66.7 95 419.9984 10000000 368.7824 190.3942516 51.6278
112.54 35 1041.6 1073.1 261.9 66.9 95 519.9981 10000000 362.197 207.5476852 57.30243
112 31.5 1419.1 1390.2 291.3 62.9 95 554.9979 10000000 362.197 187.1289499 51.66496
112.06 31.5 1419.1 1448.5 276.6 60.6 95 542.998 10000000 365.4897 182.929748 50.05059
112.12 31.5 1419.1 1415.5 273.2 62.4 95 549.998 10000000 365.4897 185.3988355 50.72615
112.18 31.5 1419.1 1489.4 282.3 63.2 95 539.998 10000000 358.9043 178.8396314 49.82934
112.24 31.5 1419.1 1219.2 282.2 62.3 95 539.998 10000000 375.3678 226.5703795 60.35957
112.3 31.5 1419.1 1498.8 318.5 58.7 95 559.9979 10000000 349.0262 173.8897123 49.82139
112.36 31.5 1419.1 1429.6 268.5 60.2 95 564.9979 10000000 365.4897 180.8847945 49.49108
112.42 31.5 1419.1 1489.4 263.7 61.1 95 409.9985 10000000 375.3678 218.6251191 58.24291
112.48 31.5 1419.1 1390.9 277.9 59.7 95 419.9984 10000000 378.6605 232.9813633 61.52777
5-21 ANN data for effect of temperature on pavement section
It can be seen that effect of temperature doesn’t give very good accuracy. Reason being
IRC analysis doesn’t take any linear or nonlinear relationship for temperature variation other
than correction for deflection values measured at pavement temperature other than 35◦C should
be 0.01mm for each degree change from standard temperature. While in use a linear relationship
is taken into account which is given earlier in theory section.
Finally, results indicate that the trained network of overlay thickness will give a quite
close approximation to the calculated values for the pavement cross sections. Accordingly, ANN
can be effectively used to determine the overlay thickness.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 49
SUMMARY
ANN has recently received lots of attention and it has also contributed in a wide variety
of applications in civil engineering as well as in other fields. In this study, ANN-based pavement
overlay design tool has been developed using MATLAB software. Several network architectures
were trained using training data sets developed by M-E overlay design program, one being used
in this study being WINFLEX (obtained through internet given in reference).
Trained network has been tested using different in field testing data sets to determine the
predicted overlay thickness, while the calculated ones have been determined by the WINFLEX
program. The calculated and the predicted overlay thicknesses have been compared together to
come up with the accuracy rate. From the sensitivity analysis and testing process using actual
data from MORTH Research scheme R-81, the average accuracy is found to be between 90 to
100% for almost all the design-cases. Other than thin and thick pavements, Cold mixed
pavement section was also used to check for usability of ANN. Effect of temperature variation
was also checked for and the accuracy of it was low due to unavailability of any linear
relationship for temperature variation in code. Finally, the results indicate that the ANN
technology can be used to predict the pavement overlay thickness with high accuracy for both
thin and thick pavements.
.
FUTURE SCOPE OF WORK
This ANN based Pavement design Procedure can be implemented in a program to get
overlay thickness as IRC based overlay thickness design procedure is yet to see dedicated
software for the same. Also dedicated MATLAB toolbox may be designed for the same.
Effect of temperature can be predicted using Linear relationship where regression linear
equation solving method may be used.
The pavement overlay design method can be extended to other types of overlay design
similar to cold mix based pavement design etc. whose straight calculation is unavailable in IRC.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 50
REFERENCES
1. AASHTO (1999),Development of a Bituminous Pavement Design Method for Europe Online,
http://ec.europa.eu/transport /road
2. Abdallah, I., Ferregut, C., Nazarian, S. and Lucero, O. M. (2000). Prediction of Remaining Life of
Flexible Pavements with Artificial Neural Networks Models. Nondestructive Testing of Pavements
and Backcalculation of Moduli, 3 rd Volume, Eds. Tayabji, S.D. and Lukanen, E. O, ASTM STP
1375, ASTM, West Conshohocken, PA, pp. 484-498.
3. Bayomy F. and Shah Z., Development of Recommendations and Guidelines for Pavement
Rehabilitation Design Procedure for the State of Idaho, Final Report 92-35-112, 1993.
4. Bayrak M., Guclu A., and Ceylan H., Rapid Pavement Backcalculation Technique for Evaluating
Flexible Pavement System, in Proceedings of the Mid Continent Transportation Research Symposium,
Iowa, pp. 319-326, 2005.
5. Burmister, D.M., The General Theory of Stresses and Displacements in Layered Soil Systems, Journal
of Applied Physics Vol/16-2,3,5, 1945.
6. COURAGE – Construction with Unbound Road Aggregates in Europe – Final Report, European
Commission – Road Transport Research, 1999.
7. Fwa, T.F., Tan, C.Y. and Chan, W.T. (1997). Backcalculation Analysis of Pavement Layer Moduli
Using Genetic Algorithms. Transportation Research Record No. 1570, Transportation Research
Board, Washington, DC, pp. 134-142.
8. Gholamali Shafabakhsh, Reza Noroozi, Fazel Fasihi (2013). Optimized ANN Algorithm for
Estimating HMA Overlay Thickness. Journal of Basic and Applied Scientific Research, 3(6)187-195
9. Gopalakrishnan K., Effect of training algorithms on neural networks aided pavement diagnosis,
International Journal of Engineering, Science and Technology Vol. 2, No. 2, 2010, pp. 83-92
10. Howard Demuth, Mark Beale, Neural Network Toolbox For Use with MATLAB, User’s Guide,
Version 4, The MathWorks, 2002
11. Huang H., Pavement Analyses and Design, New Jersey, 1993.
12. IRC: 37. (2001). Guidelines for Design of Flexible Pavements, Indian Roads Congress, New Delhi,
India.
13. IRC: 81. (1997). Guidelines for Strengthening of Flexible Road Pavements using Benkelman
Beam Deflection Technique. Indian Roads Congress, New Delhi, India.
14. Kumar, R.S., Kumar, S., Das, A., Reddy, K.S., Mazumdar, M. and Pandey, B.B. (2001).Development
of an Impact Apparatus for Evaluation of Elastic Modulus of Pavement Layers. Indian Geotechnical
Society Journal, Vol.31, No.3, IGS, New Delhi, July2001, pp. 273-284.
15. Kumar, R.S. (2001). Structural Evaluation of Pavements using Falling Weight Deflectometer. Ph. D
Thesis, Indian Institute of Technology, Kharagpur, India.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 51
16. Mahoney, J.P and Pierce, L.M. (1996). Examination of State Department of Transportation Transfer
Functions for Mechanistic-Empirical Asphalt Concrete Overlay Design. Transportation Research
Record No. 1539, Transportation Research Board, Washington, DC, pp. 25-32.
17. May, R.W. and Witczak, M.W. (1981). Effective Granular Modulus to Model Pavement Response.
Transportation Research Record No. 810, Transportation Research Board, Washington, DC, pp. 1-9.
18. Meier R. and Tutumluer E., Uses of Artificial Neural Networks in the Mechanistic Empirical Design
of Flexible Pavements, Rotterdam, 1998.
19. Mitchell, J.K. and Monismith, C.L. (1977). A Thickness Design Procedure for Pavements with Thin
Cement Stabilized Bases and Thin Asphalt Surfacings. Proceedings of 4th International Conference on
Structural Design of Asphalt Pavements, Vol. 1, pp. 409-416.
20. Mostafa Abo-Hashema (2009). Artificial Neural Network Approach for Overlay Design of Flexible
Pavements. The International Arab Journal of Information Technology, Vol. 6, No. 2 , pp. 204-212
21. Reddy, M.A., Reddy, K.S. and Pandey, B.B. (2002b). Evaluation of Pavement Layer Moduli using
Genetic Algorithms. International Journal on Pavement Engineering and Asphalt Technology, pp. 6-
19.
22. Retherford, Jennifer Queen (2012). Management of Uncertainty for Flexible Pavement Design
Utilizing Analytical and Probabilistic Methods, PhD Dissertation, Nashville, Tennessee, Graduate
School of Vanderbilt University.
23. Road Research Scheme R-56. (1999). Final report on Development of a Computer Program and
Design Charts for Analytical Design of Flexible Pavement. Submitted to the Ministry of Surface
Transport by IIT, Kharagpur, India.
24. Road Research Scheme R-81 .(2002).Final report on Structural Evaluation of Pavements in Eastern
India using Falling Weight Deflectometer. Submitted to the Ministry of Surface Transport by IIT,
Kharagpur, India.
25. Saltan M., Tigdemir M., and Karasahin M., Artificial Neural Network Application for Flexible
Pavement Thickness Modeling, Computer Journal of Turkish Journal Engineering Environment
Science, vol. 26, no. 1, pp. 243-248, 2002.
26. Seed, H.B., Mitry, F.G., Monismith, C.L. and Chan, C.K. (1965). Prediction of Flexible Pavement
Deflections from Laboratory Repeated Load Tests. Report No.Te-65-6, Soil Mechanics and
Bituminous Material Testing Laboratory, University of California, Berkeley, California.
27. Thube, D.T., Jain, S.S., and Parida, S.S. (2006). Application of Artificial Neural Network (ANN) for
Prediction of Pavement Deterioration for Low Volume Roads in India, 22nd ARRB Conference,
Canberra, Australia.
28. Venayagamoorthy, V., Allopi, D., Venayagamoorthy, G. K., 2002, Using artificial neural network as a
design tool for flexible pavement design. Proceeding of the IPET International Conference on
Engineering Technology Research, p. 58-61.
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 52
Appendix-A
A-1 EXAMPLE OF WINFLEX OUTPUT FILE
WINFLEX 2006's REPORT FOR MULTIPLE LOCATIONS 05-May-14 9:51:00 AM FILE: DESCRIPTION: ONLY FATIGUE 1. SUMMARY OF INPUT DATA -------------------------------------------------------------------------------- 1.1 TRAFFIC DATA ---------------------------------------------------------------------- DESIGN WHEEL LOAD (lb) = 4500 DESIGN DUAL TIRE SPACING (in.) = 13.5 TIRE PRESSURE (psi) = 80 DESIGN FUTURE TRAFFIC (ESALs) = 10,000,000 FATIGUE SHIFT FACTOR FOR NEW ASPHALT = 1 FATIGUE SHIFT FACTOR FOR OLD ASPHALT = 1 1.2 SEASONAL VARIATION DATA ---------------------------------------------------------------------- SUBGRADE VARIATION ------------------ No. WINTER SPRING --- ------ ------ 1 0.79 0.90 2 0.80 0.90 3 0.81 0.91 4 0.81 0.91 5 0.82 0.91 6 0.81 0.91 7 0.81 0.91 8 0.82 0.91 9 0.83 0.92 10 0.82 0.91 11 0.79 0.90 12 0.79 0.90 13 0.79 0.90 14 0.78 0.89 15 0.78 0.89 16 0.79 0.90 17 0.78 0.89 18 0.77 0.89 19 0.80 0.90 20 0.79 0.90 21 0.78 0.89 22 0.79 0.90 23 0.78 0.89 24 0.79 0.90 25 0.77 0.89 26 0.77 0.89 27 0.77 0.89 28 0.78 0.89
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 53
29 0.80 0.90 30 0.78 0.89 WINTER SPRING SUMMER FALL ------ ------ ------ ---- SUBGRADE VARIATION 1.00 1.00 BASE/SBASE VARIATION 0.65 0.85 1.00 1.00 TRAFFIC VARIATION 1.00 1.00 1.00 1.00 TEMPERATURE VARIATION(F) 44.00 58.00 66.00 36.00 PERIOD (MONTHS) 3.00 1.00 4.00 4.00 1.3 PAVEMENT DATA ---------------------------------------------------------------------- CLIMATIC ZONE : 3 (Idaho Zone) POISSON'S RATION OLD AC LAYER 0.40 BASE LAYER 0.40 SUBGRADE 0.50 SUBGRADE TYPE: LINEAR BASE TYPE: LINEAR MODULI AND THICKNESSES DATA (From ETF FILE:THICK OLD Km 123.795 to 124.000 of NH-6.etf) ---------------------------------------------- *************************************************************** No. MILE POST TEMP. (F) E1(ksi) E2(ksi) ----- --------- ----------- ------- ------- 1 124 87.8 104.6592043 30.3128793 2 123.995 87.8 107.6179734 23.30755839 3 123.95 87.8 92.89664685 26.16480108 4 123.945 87.8 122.1072396 23.84419788 5 123.9 87.8 104.3401214 23.11900938 6 123.895 87.8 116.3492429 23.61213756 7 123.85 87.8 123.9202109 29.03654754 8 123.845 87.8 125.6751671 19.94268375 9 123.8 87.8 72.57686508 30.53043585 10 123.795 87.8 110.4607123 24.71442408 11 124 77 138.6850487 35.22965733 12 123.995 77 124.1812787 36.68003433 13 123.95 77 100.1050205 33.89531049 14 123.945 77 144.3705266 55.20134862 15 123.9 77 138.481996 57.91355361 16 123.895 77 108.0965978 36.44797401 17 123.85 77 114.6377981 37.78232085 18 123.845 77 140.4690125 57.50744805 19 123.8 77 129.6492 57.01431987 20 123.795 77 78.82798995 55.72348434 21 124 105.8 75.05700975 45.12122847 22 123.995 105.8 140.0629069 29.51517195 23 123.95 105.8 95.21725005 42.75711396 24 123.945 105.8 89.0531478 36.23041746 25 123.9 105.8 91.93939803 29.66020965 26 123.895 105.8 115.8561148 46.80366579 27 123.85 105.8 117.5965672 45.77389812 28 123.845 105.8 95.04320481 40.36399191
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 54
29 123.8 105.8 94.21648992 24.71442408 30 123.795 105.8 123.6011279 31.64722614 *************************************************************** MODULI AND THICKNESSES DATA (Cont'd) ************************************************************** No. E3(ksi) H1(in.) H2(in.) ----- ------- ------- ------- 1 7.9770735 2.8851595 4.099115677 2 7.17936615 2.8851595 3.976221493 3 6.71524551 2.8851595 3.61590808 4 6.03356832 2.8851595 4.432559752 5 5.2213572 2.8851595 3.870946547 6 5.74349292 2.8851595 4.250641122 7 6.07707963 2.8851595 4.645312049 8 5.32288359 2.8851595 4.422428691 9 4.65571017 2.8851595 3.13137904 10 5.14883835 2.8851595 4.10528241 11 8.13661497 2.8851595 5.281807015 12 8.00608104 2.8851595 4.885374164 13 7.84653957 2.8851595 4.069603453 14 10.47172194 2.8851595 6.061017807 15 9.15187887 2.8851595 5.96455248 16 8.47020168 2.8851595 4.389833101 17 9.16638264 2.8851595 4.629014254 18 11.603016 2.8851595 6.012564903 19 7.42593024 2.8851595 5.668989766 20 7.9770735 2.8851595 4.086341729 21 10.08012015 2.8851595 3.649825113 22 8.62974315 2.8851595 5.150103212 23 9.47096181 2.8851595 4.190295232 24 9.05035248 2.8851595 3.804874405 25 10.63126341 2.8851595 3.692992245 26 11.12439159 2.8851595 4.939993801 27 11.08088028 2.8851595 4.961577368 28 9.47096181 2.8851595 4.112330105 29 7.16486238 2.8851595 3.611943751 30 9.97859376 2.8851595 4.714908038 ************************************************************** 1.4 OVERLAY DATA ---------------------------------------------------------------------- OVERLAY MODULUS(ksi) = 110.5144 AT TEMPERATURE (F) = 87.8 POISSON'S RATIO = 0.35 MINIMUM THICKNESS(in.)= 1 2. RESULTS -------------------------------------------------------------------------------- 2.1 OVERLAY RESULTS ---------------------------------------------------------------------- ******************************************************************************* No. OVERLAY(in.) DAMA1 DAMA2 DAMA3 DAMA4 ----- ------------ ----- ----- ----- ----- 1 15.5 0.24164 0.98129 0 0.00026 2 15.9 0.24371 0.99913 0 0.00026 3 16.4 0.26717 0.98382 0 0.00027 4 15.9 0.2222 0.97351 0 0.00029
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 55
5 16.7 0.25571 0.98612 0 0.00031 6 16.1 0.23617 0.99361 0 0.0003 7 15.5 0.22238 0.99939 0 0.00033 8 16.2 0.22407 0.9819 0 0.0003 9 17.7 0.31533 0.98748 0 0.00032 10 16.5 0.24464 0.97675 0 0.00032 11 16 0.33058 0.98231 0 0.00025 12 16.3 0.35581 0.97829 0 0.00026 13 17.1 0.40021 0.98218 0 0.00025 14 14.1 0.33261 0.98845 0 0.00025 15 14.3 0.34982 0.9979 0 0.0003 16 16.6 0.38443 0.97319 0 0.00024 17 16.1 0.38174 0.99923 0 0.00024 18 13.9 0.33768 0.97081 0 0.00023 19 15 0.36725 0.99275 0 0.00034 20 16.3 0.50072 0.97425 0 0.00031 21 12.3 0.08527 0.97643 0 0.00036 22 11.3 0.0359 0.97747 0 0.00043 23 11.8 0.06325 0.9902 0 0.00041 24 12.3 0.0736 0.98372 0 0.00037 25 12.1 0.06926 0.99239 0 0.0003 26 10.8 0.04127 0.99453 0 0.00042 27 10.8 0.04047 0.99421 0 0.00042 28 11.9 0.06416 0.98525 0 0.00039 29 12.9 0.07807 0.99436 0 0.00037 30 11.3 0.0428 0.99954 0 0.00038 ******************************************************************************* DAMA1 = FATIGUE DAMAGE ON OVERLAY DAMA2 = FATIGUE DAMAGE ON OLD AC DAMA3 = FATIGUE DAMAGE ON BTB DAMA4 = RUTTING DAMAGE
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 56
A-2 EXAMPLE OF TRAINED ANN DATASET ALONG WITH PERFORMANCE, TRAINING SET AND
REGRESSION PLOT
Considered set is CASE THICK_1: THICK OLD Km 123.795 to 124.000 of NH 6
A-1 Matlab screen shot for THICK OLD Km 123.795 to
124.000 of NH 6
A-2 Performance Data
A-3 Training State Plot
A-4 Regression Plot
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 57
Appendix B RADIAL BASIS FUNCTION DATA UNDER ANN FOR DIFFERENT DESIGNCASES
B-1 RBF data for 3layered pavement: Design-case 2
B-2 RBF data for 3layered Pavement: Design-case 3
B-3 RBF data for 3layered Pavement: Design-case 5
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 58
B-4 RBF data for 3layered Pavement: Design-case 6
B-5 RBF data for 3 layered Pavement: Design-case 7
B-6 RBF data for 4layered pavement: Design-case 2
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 59
B-7 RBF data for 4layered Pavement: Design-case 3
B-8 RBF data for 4layered Pavement: Design-case 5
B-9 RBF data for 4 layered Pavement: Design-case 7
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 60
Appendix-C
FIELD EVALUATION OF PAVEMENT USING FWD
Backcalculated Layer moduli from FWD test and Thickness details of different Stretches are given here. [MORTH Research Scheme R-81]
C-1 THIN PAVEMENTS
CASE Thin_1: Km 1.865 to 2.000 of SH * (Salua Road) for the Deflection Data Collected during the
Year 2001-02.
Thickness Details for the Stretch from Km 1.820 to 2.000 of SH*(SALUA Road):
Table C-2 Thickness Details for the Stretch from Km 1.820 to 2.000 of SH*(SALUA Road) (MORTH Research Scheme R-81)
Table C-1 CASE Thin_1: Km 1.865 to 2.000 of SH * (Salua Road) for the Deflection Data Collected during the Year 2001-02. (MORTH Research Scheme R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 61
CASE Thin_2: Km 2.850 to 3.000 of SH* (Salua Road) for the Deflection Data Collected during the
Year 2001-02.
Table C-3 CASE Thin_2: Km 2.850 to 3.000 of SH* (Salua Road) for the Deflection Data Collected during the Year 2001-02.
(MORTH Research Scheme R-81)
Thickness Details for the Stretch from Km 2.850 to 3.000 of SH (SALUA Road)
Table C-4 Thickness Details for the Stretch from Km 2.850 to 3.000 of SH (SALUA Road) (MORTH Research Scheme R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 62
Case Thin_3: Km 4.625 to 5.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-02.
Table C-5 Case Thin_3: Km 4.625 to 5.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-02.
(MORTH Research Scheme R-81)
Thickness Details for the Stretch from Km 4.625 to 5.000 of SH (IIT Bypass)
Table C-6Thickness Details for the Stretch from Km 4.625 to 5.000 of SH (IIT Bypass) (MORTH Research Scheme R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 63
CASE Thin_4: Km 3.370 to 4.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-02
Table C-7 CASE Thin_4: Km 3.370 to 4.000 of SH (IIT Bypass) for the Deflection Data Collected during the Year 2001-02
(MORTH Research Scheme R-81)
Thickness Details for the Stretch from Km 3.370 to 4.000 of SH (IIT Bypass)
Table C-8 Thickness Details for the Stretch from Km 3.370 to 4.000 of SH (IIT Bypass) (MORTH Research Scheme R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 64
CASE 5: Km 15.000 to 15.270 of NH-60 for the Deflection Data Collected during the Year 2001-02
Table C-9 CASE 5: Km 15.000 to 15.270 of NH-60 for the Deflection Data Collected during the Year 2001-02 (MORTH
Research Scheme R-81)
Thickness Details for the Stretch from Km 15.000 to 15.270 of NH-60
Table C-10 Thickness Details for the Stretch from Km 15.000 to 15.270 of NH-60 (MORTH Research Scheme R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 65
C-2 THICK PAVEMENTS
CASE Thick_1: for Km 123.795 to 124.000 of NH-6 for the Deflection Data collected during the
Year 2000-01 (Kumar, 2001)
Table C-11 CASE Thick_1: Km 123.795 to 124.000 of NH-6 for the Deflection Data collected during the Year 2000-01 (Kumar,
2001) (MORTH Research Scheme R-81)
Thickness Details for the Stretch from Km 123.795 to 124.000 of NH-6
Table C-12 Thickness Details for the Stretch from Km 123.795 to 124.000 of NH-6 (MORTH Research Scheme R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 66
CASE Thick_2: for Km 125.000 to 125.270 of NH-6 for the Deflection Data collected during Summer Season of the Year 2001-02.
Table C-13 CASE Thick_2: Km 125.000 to 125.270 of NH-6 for the Deflection Data collected during Summer Season of the
Year 2001-02. (MORTH Research Scheme R-81)
Thickness Details for the Stretch from Km 125.000 to 125.270 of NH-6
Table C-14 Thickness Details for the Stretch from Km 125.000 to 125.270 of NH-6 (MORTH Research Scheme R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 67
CASE Thick_3: Km 134.000 to 134.270 of NH-6 for the Deflection Data collected during the Year 2001-2002.
Table C-15 CASE Thick_3: Km 134.000 to 134.270 of NH-6 for the Deflection Data collected during the Year 2001-2002.
(MORTH Research Scheme R-81)
Thickness Details for the Stretch from Km 134.000 to 134.270 of NH-6
Table C-16 Thickness Details for the Stretch from Km 134.000 to 134.270 of NH-6 (MORTH Research Scheme R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 68
CASE Thick_4: for Km 150.000 to 150.245 of NH-6 for the Deflection Data collected during the Year 2001-02
Table C-17 CASE Thick_4: Layer Moduli for Km 150.000 to 150.245 of NH-6 for the Deflection Data collected during the Year
2001-02 (MORTH Research Scheme R-81)
Thickness Details for the Stretch from Km 150.000 to 150.245 of NH-6
Table C-18 Thickness Details for the Stretch from Km 150.000 to 150.245 of NH-6 (MORTH Research Scheme R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 69
CASE Thick_5: Km 151.000 to 151.245 of NH-6 for the Deflection Data collected during the Year 2001-02
Table C-19 CASE Thick_5: Km 151.000 to 151.245 of NH-6 for the Deflection Data collected during the Year 2001-02 (MORTH
Research Scheme R-81)
Thickness Details for the Stretch from Km 151.000 to 151.245 of NH-6
Table C-20 Thickness Details for the Stretch from Km 151.000 to 151.245 of NH-6 (MORTH Research Scheme R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 70
CASE Thick_7: Km 152.000 to 152.245 of NH-6 for the Deflection Data collected during the Year 2001-02
Table C-21 CASE Thick_7: Km 152.000 to 152.245 of NH-6 for the Deflection Data collected during the Year 2001-02 (MORTH
Research Scheme R-81)
Thickness Details for the Stretch from Km 152.000 to 152.245 of NH-6
Table C-22 Thickness Details for the Stretch from Km 152.000 to 152.245 of NH-6 (MORTH Research Scheme R-81)
Overlay of Flexible Pavements : An Artificial Neural Network Approach 2014
National Institute of Technology, Rourkela Page 71
CASE Thick_8: Km 153.000 to 153.245 on NH-6 for the Deflection Data collected during the Year 2001-02
Table C-23 CASE Thick_8: Km 153.000 to 153.245 on NH-6 for the Deflection Data collected during the Year 2001-02 (MORTH
Research Scheme R-81)
Thickness Details for the Stretch from Km 153.000 to 153.245 of NH-6
Table C-24 Thickness Details for the Stretch from Km 153.000 to 153.245 of NH-6 (MORTH Research Scheme R-81)